Visas and Walls: Border Security in the Age of Terrorism 9780812295696

States' migration and border control policies must enhance security while facilitating commerce. Sealing off border

127 44 2MB

English Pages 256 Year 2018

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
CONTENTS
PREFACE
Introduction: Globalization, Security, and Border Control
Chapter 1. Harder Borders in a New Security Climate
Chapter 2. Terrorism, Trade, and Visa Restrictions
Chapter 3. Terrorism, Trade, and Visa Policies in the European Union
Chapter 4. Terrorism, Trade, and Border Fences
Chapter 5. Turkey’s Migration Policy: An Illustrative Case Study
Conclusion: Improving Theory and Policy
NOTES
ACKNOWLEDGMENTS
Recommend Papers

Visas and Walls: Border Security in the Age of Terrorism
 9780812295696

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Visas and Walls

This page intentionally left blank

Visas and Walls Border Security in the Age of Terrorism

Nazli Avdan

U NI VE RS I T Y O F P E NNS YLVAN I A P R E S S P H I LADE LP H I A

Copyright © 2019 University of Pennsylvania Press All rights reserved. Except for brief quotations used for purposes of review or scholarly citation, none of this book may be reproduced in any form by any means without written permission from the publisher. Published by University of Pennsylvania Press Philadelphia, Pennsylvania 19104​-­4112 www​.­upenn​.­edu​/­pennpress Printed in the United States of Amer­i­ca on acid-­free paper 1 3 5 7 9 10 8 6 4 2

Library of Congress Cataloging-­in-­Publication Control Number: 2018032185 ISBN 978-0-8122-5105-0

For my ­family

This page intentionally left blank

CONTENTS

Preface ix Introduction: Globalization, Security, and Border Control

1

Chapter 1. Harder Borders in a New Security Climate

19

Chapter 2. Terrorism, Trade, and Visa Restrictions

40

Chapter 3. Terrorism, Trade, and Visa Policies in the Eu­ro­pean Union

78

Chapter 4. Terrorism, Trade, and Border Fences

118

Chapter 5. Turkey’s Migration Policy: An Illustrative Case Study

148

Conclusion: Improving Theory and Policy

182

Notes 201 Bibliography 209 Index 233 Acknowl­edgments

243

This page intentionally left blank

PREFACE

A book is a long-­haul effort and demands time commitment. With good reason, then, each book has a personal tale ­behind it. As a mi­grant from Istanbul, Turkey, myself, I live my research, as some of my colleagues like to remind me. I have moved across the Atlantic Ocean four times, thrice to the United States and once from the United States to the United Kingdom. I encountered the pro­cesses, paperwork, complexity, and hurdles of both short-­term and long-­ term migration. As I discuss at length in the book, in terms of pace and ease ­ ehind trade and finance. For mi­grants, borders of mobility, migration trails b are salient, despite the most optimistic pronouncements of globalists on the irrelevance of borders. For mi­g rants from some countries, borders ­matter even more. As the following chapters elaborate, scholars have observed that mobility rights have expanded but disproportionately so, by favoring a subset of countries. My own traveling and migrating several times brought this observation into sharp relief for me. Of course, I am by no means unique in my migration experience. However, it was personal experience with unequal mobility rights that narrowed my interest in borders, migration, and security into a research question. At the same time, I noted that the experience of mi­grants sharply contrasts with globalization scholars’ claims of a borderless world. Hence, although I focus on state-­level policies in the book, my interest in the subjects I address arose from the perspective of the mi­grant. This book is about transnational terrorism, globalization, and migration policies. While pundits and prac­ti­tion­ers debate what migration reforms are po­liti­cally feasible, necessary, security enhancing, and eco­nom­ically beneficial, I take an analytical approach to studying how ­t hese concerns interact to shape short-­term migration policies. Rather than proposing what constitutes optimality, I examine variation in states’ border-­and migration-­control

x Preface

policies. Some states have walls while o ­ thers have completely open borders. ­ thers restrict short-­term moSome have generous visa-­waiver programs and o bility. Equally impor­tant, some countries’ citizens face steeper hurdles in migrating, or, put differently, some passports carry mobility privileges while ­others do not. What explains this variation? That is the question that motivated this book. As this book goes to press, U.S. president Donald Trump has requested $33 billion for 316 miles of new fencing, enforcements for 407 miles of existing fence, surveillance and control technology, and training and recruitment of personnel. The fence is projected to cover 2,000 miles of the border with Mexico by 2027. The call has met withering criticism from po­liti­cal opponents while being hailed by supporters for enhancing security. Moreover, the proposed border wall figures as a bargaining chip in negotiations between Demo­crats and Republicans on immigration reform. The United States is not a sui generis example of the politicization of migration. In Eu­rope, the link between terrorism and migration has underpinned the rise of pop­u­lism, xenophobia, and the far right. The vulnerability of migration systems figured into debates over Brexit. Migration and border policies often inspire fractious debate and polemic. This is with good reason ­because, like trade and financial flows, migration brings economic consequences and distributional costs to host socie­ties. Yet unlike flows of goods and capital, flows of mi­grants trigger security fears, which typically coalesce around what it means to be a nation. Since September 11, and more so with the violent aftermath of the Arab Spring, transnational terrorism has loomed larger on the public agenda. Migration increasingly triggers fears over physical security. This has given rise to the frequently cited migration-­security nexus. As such, ­human mobility now touches three facets of security: economic, cultural, and geopo­liti­cal. The book casts light on a central dilemma countries face. A ­ fter terrorist events, governments face mounting pressures to seal off their borders. At the same time, closing off borders carries economic penalties. How do states balance the twin objectives of economic maximization and security enhancement? This is the main issue the book addresses. It does not tackle long-­term migration policies such as naturalization and citizenship rights and instead looks at how states monitor short-­term mobility. I focus on visa policies and ­ ese policies control migration before individuals have border barriers. Th crossed borders. As I discuss in the following pages, states’ migration poli-

Prefacexi

cies are interconnected. States also differ in terms of what types of migration policies become salient and require policy reforms. Hence, my book constitutes the first step in uncovering the variations in terms of visa and border policies. It also calls for ­future research to examine how globalization and transnational threats shape other aspects of migration control.

This page intentionally left blank

INTRODUC TION

Globalization, Security, and Border Control

The Question Threats to state security are sometimes carried on the backs of individuals. And once they cross borders, ­t hese threats can wreak harm on destination countries. The Christmas market attack in Berlin in 2016 is a vivid example: a mi­grant to Germany carried out a terrorist assault against German citizens, on German soil. In December 2016, a Tunisian citizen, Anis Amri, drove a lorry into the Berlin Christmas market, killing twelve and wounding forty-­ eight (Eddy 2016). One of the deadliest incidents in German history, it brought the connection between migration and terrorism into sharp relief. The ­Islamic State (ISIS) claimed responsibility for the attack. Amri was linked to another militant, an Iraqi Salafist who had also made it into Germany. The Berlin attack highlighted the fact that migration could expose the country to infiltration by foreign militants. Amri had indeed crossed international borders several times, initially reaching Italy in 2012 and then Germany in 2015. Furthermore, he had gone through Germany’s asylum system, obtaining papers to remain in the country. Not surprisingly, critics of Germany’s welcoming refugee policies cast blame on Chancellor Angela Merkel, who had in the previous year opened the country’s borders to half a million mi­grants. A string of terrorist attacks in recent years has fueled the debate over terrorism and migration control. With each fatal attack on Eu­ro­pean soil, the migration-­terrorism linkage has become more pronounced and states have been quicker to exhort policy stringency. Terrorist events have been publicized widely and the media has been quick to draw attention to target states’ demands for tighter control over borders. ­A fter the Brussels attacks in March 2016, the Eurasia Group released a report noting, “Combined, t­ hese

2 Introduction

attacks w ­ ill increase xenophobic and anti-­immigration sentiment across the E.U., which has already been rising in light of the E.U.’s ongoing refugee crisis” (Erlanger 2016). The attacks gave ammunition to right-­w ing leaders in Eu­rope to demand draconian policies. For example, ­after ISIS coordinated attacks in Paris, France’s far right leader, Marie Le Pen, urged Eu­ro­pean states to abolish freedom of movement, a cherished pillar of the Schengen regime: “Without borders, neither protection nor security are pos­si­ble” ­ fter the March  22 attacks in Brussels, (Troinanovski and Walker 2015). A Mike Hookem, a member of the Eu­ro­pean Parliament from the UK In­de­pen­ dence Party, declared that the “horrific act of terrorism shows that Schengen ­free movement and lax border controls are a threat to our security” (Erlanger 2016). The common thread in the responses to the attacks was that unmonitored flows of mi­g rants are dangerous and that transnational terrorism ­warrants greater control over borders. Clearly, for some Eu­ro­pean leaders, the knee-­jerk reaction to ­these attacks was to emphasize the dangers of migration—­terrorists gain access to Eu­ro­ pean states’ soil—­and to then call for tighter border controls. The Christmas market attack vividly exemplifies the connection between transnational terrorism and border and migration policies. Border crossing was at stake in other attacks in Eu­rope, and of course countries outside of Eu­rope have suffered terrorist vio­lence perpetrated by mi­grants. The Reina nightclub attack on January 1, 2017, in Istanbul, for example, was staged by an Uzbek national. One of the 2013 Boston marathon bombers was a Chechen, and the other was a naturalized American. Regardless of their immigrant status, the perpetrators had crossed borders and migrated to the United States. Th ­ ese events stirred latent fears over lax border control and the vulnerability of migration systems. Increasingly in U.S. politics, migration and border control have occupied the headlines. The campaign trail to the November 2016 presidential election was littered with polemic and controversy over the need for harder borders. The Trump administration spelled a restrictionist bent in migration policies. Since taking office, Trump has barred travelers from select Muslim-­majority countries, limited the number of refugees from certain countries, increased immigration arrests and deportations, and proposed a controversial bill aiming to drastically reduce ­legal migration (Baker 2017). In 2017, Trump reiterated his campaign pledge to build a border wall and to have Mexico pay for it. Shortly ­after that, he proposed tougher mea­sures for screening green-­ card applicants (Kopan 2017). Trump often justifies his demands for a com-

Introduction3

prehensive tightening of migration and stepped-up border control with references to the terrorism-­migration nexus. We would thus conclude that the specter of terrorism ­w ill orient states ­toward harder borders. We might expect transnational terrorism to universally spur states to adopt tighter policies. Countries vary tremendously in how they control their borders. Some countries have walls; ­others are calling for walls. Some countries have open visa policies and ­others do not. What explains the variation in how states monitor their borders? This is the subject of my book. Do security concerns drive border closure? Speculation abounds on this question, but thus far scholars have not marshaled hard evidence to explain the complexities of this relationship. I argue that the answer to this question is not a straightforward yes. For one, not all terrorist events are fatal, and even when they are, they do not elicit the same reaction. For another, very rarely do countries enact w ­ holesale border closure. Border closure is a misnomer ­because border control is multifaceted. Put succinctly, “immigration and border policy is much more nuanced than terms such as ‘open’ or ‘closed’ can capture” (Rudolph 2006, 27). Also, globalization further complicates the situation. On the one hand, countries need open borders for trade and f­ ree flow of workers. On the other hand, open borders have been heavi­ly criticized for leaving countries open to terrorism. Fortunately, states have multiple tools at their disposal with which to control borders. How politicians manipulate ­t hese vari­ous border controls is a careful balancing act between the state’s economy and security. Two of t­ hese tools are border walls and visa policies.1 What both instruments share in common is that they allow states to monitor migration upstream—­t hat is, before individuals have crossed borders. In contrast, naturalization and citizenship policies control migration downstream—­after individuals have crossed borders (Meyers 2000). Naturalization and citizenship policies deal with longer-­ term migration while visas and barriers regulate short-­term migration. The scholarship on policies for short-­term migration is rather thin (Mau et al. 2015). This scarcity of research is all the more glaring when we consider the connections between short-­term and long-­term migration policies (Koslowski 2009). Visa controls significantly influence longer-­term migration patterns, including asylum applications and settlement practices (Czaika and Hobolth 2016). Visa policies and border barriers differ on one significant component: visibility. Visa policies are formulated ­behind the veil of bureaucracy and are not directly observable by the public. A country’s citizens are often unaware

4 Introduction

of which states their own government grants visa waivers to. In contrast, border barriers are readily observable by the public and are high profile. They afford symbolic value, in­de­pen­dent of their objective effectiveness in forestalling illegal entry.2 Andreas (2009) contends that high-­profile enforcement initiatives are a form of security theater whereby governments demonstrate control without needing to prove the effectiveness of policies. Installing fences are about the spectacle of state authority through which governments signal that the state can defend its borders. Insofar as border instruments perform dif­fer­ent functions, we cannot anticipate that states ­will stiffen policies across the board. Given ­t hese differences, the balance between economic and security incentives ­w ill depend on the functions that border instruments serve. The trade-­off between security and economic gain has come to light time and again ­after fatal transnational terrorist events. Politicians face mounting pressure to seal borders ­after attacks. ­After the Brussels attacks, for example, criticism turned to the Schengen system, which allows individuals to cross borders within the Schengen bloc of twenty-­six countries without passport checks or immigration controls. Some commentators wondered if it should be scrapped and replaced with a tougher system. Despite pressure to seal off borders, politicians must also bear in mind economic security. Responding to calls to reinstate border checks within the Schengen zone, Ian Bremmer, Eurasia Group’s president, stressed the pos­si­ble ill effects of such a move: “­There’s g­ oing to be a lot of social instability that comes with that. It ­will tear at the fabric of what we think Eu­rope is. It’ll certainly hurt the economies.” On the flip side, Bremmer further noted, “With Paris, suspending Schengen became a question of national security. Brussels fits into this latter category” (Wang 2016). Politicians are stuck between a rock and a hard place: they know that the economy depends on open borders, but this can create insecurity and increase the risk of terrorism. This leads to the second question that animates the book: How do states balance economic and security interests when crafting border-­control policies? My argument is three pillared. First, I distinguish between global and directed (or targeted) terrorist threats. Terrorism does not translate into blanket policy tightening ­because threats elsewhere and against ­others do not ignite fears as acutely as do threats on the state’s own soil and against its own citizens. Second, economic interdependence offsets pressures ­toward tighter policies. States face the opportunity costs of diminished commerce resulting from tighter policies and fear backlash from partners. Liberal lobbies further stay the government’s hand when it comes to tougher controls. Third, policy change is not uniform ­because dif­fer­ent border-­control instruments

Introduction5

have varied functions. Take border barriers, for example. Walls assuage fears ­because they are overt manifestations of state power. They also do so by reassuring domestic audiences that the state can defend its borders. In comparison, visa restrictions may fly u ­ nder the radar, u ­ nless the government enacts them with fanfare. If such differences exist by virtue of how a par­tic­ u­lar state chooses to control its borders, we cannot always expect terrorism to predict stiffer policies across the board.

The Argument Terrorist vio­lence does not always result in harder borders. I argue that security incentives dominate policymaking when targeted attacks are the issue. In contrast, economic interdependence effectively tames fears over security when attacks are global. In the face of foreign attacks, economic interests hold sway over policymaking and make for open borders. In order to unpack this argument, I emphasize that attacks do not trigger the same level of response across all states. Some attacks ­matter more than ­others. Much has been written about the discrepancy in the public response to international terrorist events that took place in 2015. Attacks on Eu­ro­pean soil received greater attention than did attacks in ­Kenya and Turkey. Whereas Eu­ro­pean countries lit up their national monuments in the colors of the flags of France and Belgium a­ fter the bombings in ­those countries, terrorist events elsewhere did not evoke the same level of outrage (Ahmed 2016). Commentators speculated that the discrepancy resulted from an empathy gap whereby ­people identified with victims they viewed as similar. This affinity functioned regardless of geo­graph­i­cal distance. For instance, from the perspective of the United States, the London tube bombing on July 7, 2005, was more threatening than the Bali attacks in the same year ­because the latter incident was culturally distant to U.S. interests. The contrasting public responses to international events also clue us in on a simpler distinction: that between directed and diffuse threats. Vio­lence that is geo­graph­i­cally or culturally proximate strikes closer to us. More s­ imple than that, however, attacks in our own backyard and against our compatriots strike closest to us. Terrorism is more likely to produce harder borders if the state’s own interests are involved, that is, if its own nationals are hurt in attacks or if terrorists execute attacks on the state’s own territory. Thus if threats strike close to home, they more readily effectuate policy change.

6 Introduction

Terrorist events sow fear and are effective to the extent that they do so. In the absence of fear, vio­lence would not elicit a response from targeted states. Fear affects ­whether and when governments respond and the shapes that responses take. Threat perception forges the implicit link between terrorist event and policy response (Rudolph 2006).3 Along t­ hese lines, scholar­ ship has shown that public perceptions of terrorist vio­lence affect a range of policy outcomes (Huddy et al. 2002; Huddy et al. 2005). Public attitudes ­toward terrorism affect who comes to power in democracies and, more generally, reshape the po­liti­cal and social climate. Heightened threat perceptions generate support for right-­wing parties, who may then enact more draconian policies. ­These perceptions also correlate with intolerance ­toward minorities, paving the way to restrictions on civil liberties and the rights of foreigners (Peffley, Hutchison, and Shamir 2015). If we apply ­t hese insights to border-­ and migration-­control policies, several effects on policy stand out. First, heightened threat perceptions w ­ ill translate into greater support for tougher border policies. Scholars have shown that fears and anx­i­eties surrounding ­terrorism ratchet up support for harsher counterterrorism policies. This support is likely to accompany toughness on border and migration control, especially insofar as controlling borders is couched in counterterrorism terms. Second, electorates may bring to office candidates who are e­ ager to toughen border control. In addition, a fearful public grants more leeway to politicians with a tougher agenda. How terrorism impacts threat perceptions thus affects the shape that border and migration policy takes. Terrorist events heighten threat perception to differing degrees. Quite intuitively, we would expect terrorist vio­lence occasioned on a country’s own soil to ­matter more than events that transpire abroad. We might also expect threats to be more salient if a country’s own nationals are hurt. To assess t­ hese expectations, I differentiate between global and targeted (or directed) threats. Targeted threats directly imperil state interests by victimizing its nationals and endangering territorial integrity. Attacks on the state’s soil showcase that violent non-­state actors can not only cross borders unnoticed but also mobilize and launch attacks within its territory. Transnational terrorist events that transpire on other countries’ soil but hurt the state’s own nationals also incentivize policy stringency. Terrorist incidents executed on a country’s territory or that involve its citizens, even when they occur abroad, strike at the core of state sovereignty. ­Table 1 synthesizes ­t hese insights, posing threat perception as an intermediary link between the terrorist threat and policy impact. Assaults that

Introduction7

­Table 1. Terrorist Events, Threat Perception, and Policy Impact Venue

Victims

Threat Perception

Policy Impact

Global terrorism

Foreign victims Compatriots

Low Moderate to high

Muted Moderate

Targeted terrorism

Foreign victims Compatriots

Moderate to high High

Moderate Acute

victimize citizens and/or transpire on the state’s own territory are both targeted against state interests and expected to have a more pronounced impact on policies. Global attacks are defined as events that transpire abroad and do not involve the state’s own citizens. Such attacks are diffuse and remote; they do not stoke anxiety as cogently as do targeted terrorist events. We also expect attacks abroad to have a moderate impact on policies insofar as personal proximity trumps geo­graph­i­cal distance. Incidents within the state’s borders sometimes only involve foreigners, but ­these events are likely to impel tighter policies simply ­because of physical proximity. Contrarily, foreign attacks that victimize ­others’ nationals are geo­graph­i­cally and personally remote. Why ­doesn’t terrorism always produce harder borders and tougher migration polices? ­After all, terrorism is designed to strike fear into the hearts of an audience broader than the immediate victims (Hoffman 1998). By extension, transnational terrorism terrorizes not just the victims or even the targeted populace but the global audience. We might then anticipate that countries gradually toughen policies as they witness bombs go off on other shores. The answer is ­simple. The effects of global terrorism—­v io­lence on ­ eoples—­are modest relative to directed terrorother shores and against other p ism. This makes room for economic concerns to dominate policy. ­Globalized states selectively guard their borders. They prefer open border policies, as long as threats remain remote. Sealing off borders may enhance security, but it is costly. Consider visas, for example. Governments use visa requirements to monitor flows and deter unwanted travelers. The flip side, however, is that visa restrictions impose burdens on legitimate business and dampen tourism (Neumayer 2011). The Cato Institute, a libertarian think tank, estimated that visas dampen bilateral trade and foreign direct investment (FDI) by 25 ­percent. The same study also added that the United States would stand to gain $90–123 billion in annual tourist spending if it eliminated all travel visas (“Eu­rope’s Response to

8 Introduction

the Paris Attacks Is Dif­fer­ent This Time” 2015). Consider border fences. The Secure Fence Act of 2006 authorized 700 miles of fence to be built along the U.S. border with Mexico. Customs and Border Protection (CBP) spent over $2.4 billion between 2006 and 2009 on 670 miles of the U.S.-­Mexico fence. The United States spent $341 million in 2017 just to maintain the existing fence. And in September 2017, President Trump asked Congress to allocate an additional $1.6 billion ­toward building a 1,900-­mile wall (Nixon 2017). Costs are not restricted to the ­actual price tag on the border instrument. Stringent border controls “place governments on a collision course with easy trade, which is central to sustained expansion and integration of the global economy” (Flynn 2000, 58). Stringent controls run ­counter to liberal princi­ ­ fter 9/11, ples that are integral to reaping the benefits from benign flows. A for example, the United States pushed forward illiberal mea­sures that limited the rights of foreigners and immigrants (Flynn 2003). The Patriot Act is a case in point. Th ­ ese mea­sures ­were controversial not only ­because they defied the tenets of liberalism but also ­because of their deleterious economic effects. Illiberal mea­sures may impede desirable migration flows by undercutting incentives to migrate or by pushing existing mi­g rants to exit the country (Czaika and Haas 2013). Stiffer policies are also not feasible long term ­because of downstream and lasting economic losses. Shortly ­after the 9/11 attacks, the United States toughened its border-­control efforts. As a direct consequence of border closure, cross-­border traffic in the area covered by the North American F ­ ree Trade Agreement (NAFTA) slowed to a crawl in the weeks following the September 11 tragedy. The United States faced immediate economic losses from its border crackdown. Th ­ ese losses resulted directly from the logistical hurdles of border closure. ­There are also downstream and lasting economic repercussions. Such mea­sures can put a dent in cross-­border exchange by temporarily relocating trade. They can also anger commercial partners and hurt diplomatic ties and even elicit overt backlash from trade partners. In sum, at the dyadic level, we expect economic interdependence to render states reluctant to close their borders.

Globalization and Borders The foregoing discussion underscores that globalization engenders open borders. Globalists take the argument further, however, by proclaiming that globalization w ­ ill wipe out international borders (Ohmae 1990, 18). Some

Introduction9

scholars even claim that cross-­border mobility signifies the retrenchment of state powers and the weakening of territorial authority (Sassen 1998; Strange 1997). While globalists’ assertions of a borderless world seem exaggerated, ­t here is some evidence of freer movement across borders. ­There is support for both sides of the debate on the costs and benefits of maintaining borders in a globalizing world. Since World War II, more and more states’ citizens enjoy visa-­free travel. Mau et al. (2015) refer to the visa-­waiver programs as the global mobility regime. The global mobility regime is certainly expanding, but it’s also lopsided in privileging a select number of countries. While only 20 ­percent of the world’s population benefited from visa-­free travel in 1968, in 2010, that number surpassed 35 ­percent. Global migration has also risen considerably, reaching 244 million mi­grants in 2015, a 41 ­percent increase ­ ese numbers seem impressive. Nevertheless, the world’s popufrom 2000. Th lation grew at a faster rate than that of the mi­grant population. Hence, the global mi­grant pool as a ratio of the world’s population is still modest. In fact, the ratio has remained fairly steady since the 1960s, hovering at around 3 ­percent (Czaika and Haas 2015). Trade and financial flows have grown more rapidly than l­abor flows. States’ policies partially explain the lag: we know that even eco­nom­ically liberal states retain restrictions on l­egal migration (Hollifield 1992, 516). Moreover, when states have loosened controls, they have done so selectively. Mobility rights have increased for a subset of countries—­primarily citizens of advanced Western democracies (Mau et al. 2015; Neumayer 2006). Western democracies tend to have disproportionate passport power in terms of the number of countries their citizens can travel to without a visa. When it comes to their migration policies, however, Western states are not the most open. For instance, the United Kingdom, Germany, and Canada rank among the top three countries in terms of passport power; however, ­these same countries are not among the most liberal in terms of their visa policies ­toward other countries.4 The limited scope of the global mobility regime means that aside from a select subset, countries still encounter steep visa hurdles. In other words, borders continue very much to ­matter where ­human mobility is concerned: “Anyone who thinks differently should try landing at a Sydney airport without an entry visa or go to France and apply for a job without a work permit” (Freeman 1998, 93). Perhaps more vivid evidence against globalists’ prognostications of a world without borders is the recent trend t­ oward fencing borders. Since the fall of the “iron curtain,” new border barriers have cropped up, and they have

10 Introduction

done so at an accelerated pace. From 1945 u ­ ntil 2013, states built sixty-­two new fences, and they built forty-­eight of them a­ fter the end of the Cold War. More recently, the war in Syria and resulting outpouring of refugees into ­Eu­rope have spurred EU member states to fortify their borders (Batchelor 2015). As a result, non-­EU citizens now face tougher hurdles to enter the Eu­ro­pean Union. This has led to concern that Eu­ro­pean consensus on the freedom of movement is beginning to show cracks. While terrorism is not the sole reason why states erect border barriers, fears over security in general certainly serve as a po­liti­cally expedient justification for building walls and fences. President Trump, for example, lists a host of security threats including crime, narcotrafficking, and illegal migration in pushing for the construction of a wall along the border with Mexico. Yet, in most cases, politicians sincerely believe in the necessity of a fence as well as in the effectiveness of a barrier in blunting the terrorist threat. Demo­cratic leaders tout fences as security enhancing and necessary for the protection of the country against external threats. In no other context is the security link more pronounced as when politicians can point to terrorism as a concrete and formidable threat and the border wall as a panacea against that threat (Jones 2012a, 2012b). For example, Israeli prime minister Benjamin Netanyahu stressed concerns about ISIS infiltration of Jordan when trumpeting a new fence along Israel’s border with Jordan. Netanyahu underscored that the wall is necessary, noting, “We must be able to stop the terrorism and fundamentalism that can reach us from the east at the Jordan line and not in the suburbs of Tel Aviv” ­(“Netanyahu” 2014). In sum, globalists’ claims that we are headed t­ oward a borderless world ­ atter less for some types of flows, do not find much support. Borders may m but for ­people, they continue very much to ­matter. We may speculate that non-­state threats, or the “globalization of informal vio­lence,” is the culprit (Keohane 2002). Globalization has negative externalities such that the transport and communication technologies that facilitate benign flows empower threatening actors as well. Globalization increases states’ vulnerability to non-­state threats. Just as goods and money are now able to traverse larger distances in a shorter amount of time, so are individuals. If threats are carried on the backs of individuals, then more remote threats have the potential to traverse greater distances to reach target states. With globalization, perforated borders stand to endanger states’ security. Non-­state actors can wield large-­scale vio­lence, previously the preserve of states’ militaries. Th ­ ese actors

Introduction11

can surmount the gap in capabilities they face vis-­à-­v is states through surprise, secrecy, and shock. The growing menace of non-­state actors demands that we rethink our assumptions about geo­graph­i­cal space as barriers. Coincident with the rise of non-­state threats, globalization also makes pos­si­ble dangerous alliances, for example, as evidenced by the crime and terrorism nexus (Dishman 2005). As non-­state groups forge links across borders, they are further able to capitalize on globalization. Thus it is not surprising that states lean t­ oward restrictionism when it comes to ­human mobility. The preceding discussion leads us to expect that border controls are ­here to stay, even among liberal states. ­These trends may only lend circumstantial evidence on the domineering influence of security fears, however. In order to gain granular traction on how states balance economic and security objectives, we need to narrow the analytical focus by examining the impact of a specific type of security challenge: transnational terrorism. The impact of economic objectives may remain unclear, as I argue, if we look at states’ overall economic openness. Although the Eu­ro­pean Union has been castigated as “fortress Eu­rope” (Finotelli and Sciortino 2013), economic liberalism is its mainstay. Hence, it is misleading to expect economic liberalism as a philosophy, or economic openness as ­grand strategy, to directly influence migration-­and border-­control policies. To better capture the effect of economic incentives, we need to look at ties between states and analyze how ­t hese ties in turn affect policies t­oward commercial partners. This perspective takes inspiration from economic interdependence theory in arguing that economic ties affect how states seek security.

Existing Explanations of Migration and Border Policies Previous studies have not adequately analyzed the trade-­offs between trade and security with borders while also accounting for the differences in policy options for border control. My work is the first to do both of t­ hese ­t hings at the same time and evaluate hypotheses based on this framework on a ­grand scale. Migration policy remains woefully undertheorized in international relations (IR) scholarship. Where theoretical accounts exist, the bulk of the lit­er­a­ture studies migration flows rather than policies. When scholars have turned attention to migration policies, they have analyzed the policies of single countries rather than adopting a comprehensive framework. Existing

12 Introduction

work employs a case study design, engages in discourse analy­sis, or is normative in orientation. Discourse analy­sis provides an incomplete picture insofar as ­there is a gap between rhe­toric and practice. The prevailing ­approaches to migration policy form a comparativist lens, highlighting its domestic under­pinnings. A cursory survey of the lit­er­a­ture shows that ­economic and cultural arguments dominate scholarship. Comparative po­liti­cal economy informs us that firms generally support open migration policies while ­labor does not. Domestic co­a li­tions on both sides form in response to the fiscal and distributional economic effects of migration. A popu­lar perspective draws on the Heckscher-­Ohlin framework to argue that the o ­ wners of scarce resources stand to lose from open policies and thus ­w ill trumpet protectionism (Hiscox 2006; Scheve and Slaughter 2001). A related perspective refines the f­ actor endowment model to maintain that ­owners of mobile resources ­w ill advocate economic openness. At the center of both perspectives is the view that commerce bifurcates society, si­mul­ta­neously creating protectionist and liberal stakeholders (Hollifield 2000). Liberal lobbies anticipate revenue loss in the event of retaliation by trade and financial partners. Accordingly, winners from economic liberalization champion sustained liberalization and open-­border policies. Migration policy may be conceived as client politics whereby the openness of migration policy depends on the relative strength of ­these ­lobbies (Freeman 2001). Liberal lobbies have an orga­nizational advantage ­because the benefits of open migration are concentrated, but the costs are diffuse. In contrast, opposition to migration is omnipresent, yet poorly or­ ga­nized. Additionally, to the extent that trade and migration are complementary (Rudolph 2008), we would expect pro-­trade lobbies to join forces with ­t hose in ­favor of open migration policies. As such, anti-­immigration lobbies find it more difficult to mobilize and pressure the state to maintain protectionism (Freeman 1995). Insofar as liberal lobbies carry the day, we would expect the domestic-­level mechanisms to support the state-­level ­argument. The economic perspective is strictly premised on l­abor market effects of migration and ignores that unlike other f­ actors of production, l­abor comes with sociopo­liti­cal consequences (Zolberg 1987). This insight has inspired a separate strain of comparative lit­er­a­ture, which draws on the cultural effects of migration to explain policies. Migration may stoke cultural insecurities by undermining societal cohesion and challenging the bound­aries of what

Introduction13

defines a nation (Teitelbaum and Weiner 1995; Waever et al. 1993). Particularly if mi­grants carry starkly dif­fer­ent normative templates, their codes of conduct, belief systems, and self-­identities may clash with the host state’s way of life (Collier 2013). Homogeneous host socie­ties find accommodating cultural diversity particularly challenging, also ­because for ­t hese countries, the concept of nationhood is ethnic rather than civic. Such socie­ties are hermetic, inflexible, and predisposed to opposing cultural diversity. In contrast, settler states built on significant waves of immigration at historical junctures—­ such as Canada, the United States, the United Kingdom, New Zealand, and Australia—­base citizenship on civic princi­ples and hence have an easier time reckoning with dissimilar migration (Brubaker 1992). The cultural perspective thus expects countries to pursue restrictive policies in an effort to preserve a coherent sense of nationhood or, more generally, to buoy cultural security. It also maintains that states that are more ethnically homogeneous and states that lack a tradition of migration are inclined ­toward stricter ­migration and border controls. Arguments drawing on comparative politics to explain why states pursue open or closed migration and border-­control policies illuminate the domestic under­pinnings of policies. However, they cannot tell us much about how interstate dynamics affect migration policies. For example, we know from comparative po­liti­cal economy which actors forge pro-­and antimigration lobbies. But we do not know if ­t hese actors compel governments to attune liberalization ­toward economic partners. This is where theories about bilateral trade and commercial dependence have merit. From a comparative economic perspective, closed migration policies are rooted in the domestic economic consequences of l­abor flows. This perspective does not shed light on how economic ­factors interact with other facets of state ­grand strategy. As part of its g­ rand strategy, a state conjointly seeks economic, geopo­liti­cal, and cultural security (Rudolph 2005). The cultural perspective tells us that nativist lobbies exploit ethnocentric sentiment to fuel cultural insecurities ­ eople against open migration policies. But we know less about and galvanize p ­whether the economic ties between states temper domestic cultural insecurities. Moreover, the subfield of IR has rich scholarship on territoriality and borders and a wealth of theories to draw upon in tying interstate dynamics to migration policies. Yet it has not adequately leveraged this lit­er­a­ture (Gavrilis 2008a; Rudolph 2003). IR scholars’ oversight is surprising in light of the fact that international migration has assumed center stage in policy debates

14 Introduction

in the past few years. As Adamson stresses, “International scholars and ­policy makers are finding it increasingly difficult to ignore the relationship between migration and security in a highly interconnected world defined by globalization pro­cesses” (2006, 165–167). Hence, the topic of migration control provides fertile ground for theorizing from the international relations lens. ­There are a handful of scholars who have recognized the scholarly void and taken a comprehensive perspective. Rudolph (2003) argues that trading states’ migration policies tend to be more liberal given external geo­ po­liti­cal threats, as was the case during both world wars. The relationship reverses in the absence of external threats, where the absence of a common ­enemy means states can afford tougher policies. Rudolph’s work offers one perspective on when economic interests m ­ atter the most: when geopo­liti­cal threats compel states to value economic security and trigger a rally effect, thereby overriding concerns over cultural insecurity. Peters (2017) offers another perspective. She argues that globalization initially predicted open migration policies ­because of strong pro-­m igration lobbies. Deindustrialization and offshore mobility of firms shrank the pro-­immigration lobby. She maintains that ­these ­factors are why globalized states now pursue strict migration policies. Both perspectives highlight that material considerations ­matter. For Peters, the current wave of globalization fails to exert a ­ ecause the domestic pro-­migration liberalizing force on migration policies b lobby shrank as states moved production offshore and technological development blunted the demand for foreign l­abor. For Rudolph, globalization may fail to produce open policies in times of relative peace—­absent an external common aggressor. In this scenario, threat perceptions turn inward and host states become preoccupied with cultural rather than economic insecurity. Rudolph and Peters show that globalizing has varied effects on states’ migration policies. But neither work tells us much about how non-­state threats—­and informal vio­lence wielded by t­ hese actors—­influence migration-­ control policies. My argument points to a dif­fer­ent reason why economic incentives may have limited impact on open-­migration policies. Economic models stipulate that states engage in cost-­benefit analy­sis when deciding how open borders should be. Fears affect policy through an appeal to emotion rather than strict cost-­benefit analy­sis. This suggests that states deviate from strict cost-­benefit analy­sis when policies then reflect ­t hese fears (Friedman

Introduction15

2011; Mueller 2005). Coupled with weak lobbies, emotional decision making should further winnow support for open migration.

Theoretical Anchor My book speaks to debates about the role of borders and territoriality in states’ pursuit of sovereignty. Territoriality and borders occupy a venerable and impor­tant role in IR scholarship; this lends a solid theoretical foundation when generating empirical implications about borders. Nonetheless, when IR scholars have studied borders, they have done so by looking at the role of borders in conflict (Huth 1996) or cooperation (Simmons 2005). Conflict scholars examine why states clash over the location of borders. Cooperation scholars explore how borders can function as effective institutions facilitating cross-­border trade. Neither perspective tells us much about how states manage their borders. Common to both perspectives is that territorial demarcation is a given. Agnew refers to the unquestioning ac­cep­tance of demarcation as the “territorial trap,” noting that mainstream IR “assumes implicitly that a state is a fixed territorial entity operating much the same over time and irrespective of its place within the global geopo­liti­cal order” (1994, 54). The territorial trap handicaps our understanding of how states manage borders in the face of non-­state threats. An impor­tant consequence of ­globalized informal vio­lence is the emerging role of borders as perimeters of ­defense against intrusion by non-­state threats. While scholars underscore that non-­state actors can exact considerable damage (Keohane 2002; ­Salehyan 2008b), scant attention has been paid to ­whether t­hese concerns do in fact alter how states manage their borders. By looking at transnational terrorism, I shift attention to how states manage their borders in the face of such threats. This book also asks scholars to reconsider territoriality in international ­relations. The territorial trap is underpinned by realism’s contention that sovereignty is predicated on territorial exclusion. As an early voice in classical realism, Morgenthau defined sovereignty as “supreme ­legal authority of the nation to give and enforce law” (Morgenthau 1978, 4). Eighteenth-­and nineteenth-­century geopo­liti­c al thinking mirrors realist thinking and stresses that geography plays a pivotal role in state security (Starr 2006). Within its borders, the state monopolizes the legitimate use of force. Borders

16 Introduction

delimit the bounds of states’ policing and lawmaking authority. In a similar spirit, the modern state has appropriated the legitimate means of movement (Torpey 2000a; Torpey 2000b). Dating back to the turn of the twentieth ­century, the passport is a relatively new invention. In previous centuries, individuals could traverse bound­aries unencumbered by the need for documentation. The passport and the visa regime allow the modern state to monitor movement across its borders (Salter 2003). The realist perspective contrasts sharply with liberalism on the role of borders. Liberals view borders primarily as institutions that regulate and facilitate cross-­border ­exchange (Simmons 2005). Whereas realists emphasize territorial sovereignty, liberals stress managing and monitoring flows across borders, or what Krasner (1999) has dubbed interdependence sovereignty. Territorial sovereignty requires the exclusion of external actors; in contrast, interdependence sovereignty turns on “the ability of public authorities to control transborder movements” (Krasner 1999, 9). The book’s analy­sis also sheds light on a second debate between globalists and security scholars. To summarize the debate, the material benefits of transnational exchange in an increasingly globalized economy have led many observers to anticipate the erosion of state territorial control and ultimately the obsolescence of state borders. The security externalities of globalization, however, point to the opposite expectation: harder borders. To date, no one has attempted to reconcile t­ hese propositions. This debate remains unresolved also ­because it ultimately rests on an empirical question of how globalization affects border politics. The book addresses this controversy by expanding economic interdependence theory beyond its conventional bound­aries, which are limited to questions of conflict and cooperation. An extensive scholarship addresses how economic incentives can restrain states from using militarized force (Oneal and Russett 1997; Oneal and Russett 1999; Russett and Oneal 2001). We know much less about how economic ties influence how states seek security when faced with nontraditional threats. Border and migration control is an ideal testing ground for studying how economic and security incentives influence policies ­because this is one policy domain where both types of incentives intersect (Rudolph 2005). In contrast, for example, other policy domains may be about maximizing primarily one objective. Trade policy concerns economic security. Counterterrorism is about physical security. Migration and border control are about both in that states si­mul­ta­neously pursue economic and security objectives in their ­grand strategies.

Introduction17

Globalization scholars assert that commerce inspires policy liberalization ­ ecause economic interests override security incentives in state ­grand stratb egy (Rosecrance 1986). Some might go even further to say that globalization imposes a golden straitjacket on states, rendering unilateral policies infeasible (Friedman 1999). This perspective is limited, however, ­because it does not account for economic ties between pairs of states but rather expects trading states to have open borders. This limitation underwrites the enduring clash between the expectations of globalization and security scholars. Economic interdependence theory suggests that economic ties between states constrain policies. The constraining effect grows with the relative strength and salience of such ties. I propose a similar argument and maintain that asymmetric interdependence in par­tic­u­lar creates a push for softer policies ­toward trade partners. To illustrate, Canada and Mexico are disproportionately dependent on commerce with the United States; bilateral trade comprises a higher proportion of overall trade and of gross domestic product (GDP) for both states than it does for the United States. This asymmetric dependence compelled Canada and Mexico to oppose a border clampdown in the wake of 9/11. Commerce also creates stakeholders that advocate for open ­ ere vocal in opposing borders. Transnational firms and commercial interests w the thickening of borders within NAFTA. ­These debates anchor the broader implications of the book’s three interrelated findings. First, I show that states harden borders against origin countries whose nationals have conducted attacks against their own interests in the past. This is a targeted rather than universal policy response; in other words, states targeted in terrorist attacks do not exhibit a blanket response but tighten up policies against only a subset of countries. Second, trade and capital ties render states less ­eager to harden borders against the citizens of commercial partners, regardless of previous attacks. Third, the identity of victims ­matters in conditioning states’ policy choices. States ­factor in attacks against other similar countries and treat ­t hese events as striking closer to their own interests. This effect is evident in the Eu­ro­pean Union whereby attacks on the Eu­ro­pean continent compel states to tighten up their border-­ and migration-­control policies, regardless of ­whether their own citizens ­were hurt in such attacks. Th ­ ese findings yield support for both sides of the debate. I argue that borders continue to ­matter, but they take on the role of a shield against transnational threats rather than traditional, military threats. The argument refines realist insights on borders as a hard shell against the militaries of other states. My argument also upholds the liberal perspective,

18 Introduction

however, by showing that economic ties soften policies and reduce the likelihood of fortified borders. I show how the insights from economic interdependence encompass state be­hav­ior beyond the decision to engage in militarized disputes. The book demonstrates that parallel to their effects on conflict be­hav­ior, material incentives shape how states cope with transnational terrorism. The insights in this book would thus appeal to scholars interested in how security and economics intersect. The empirical results should cause students of economic interdependence theory to be cautiously optimistic: as long as vio­lence does not transpire in the state’s own backyard, commerce prohibits policy tightening. However, even distant terrorist events that harm the citizens of the state inspire policy tightening. In other words, they are regarded as occurring in the proverbial backyard.

CHAPTER 1

Harder Borders in a New Security Climate

Faced with the exigencies of globalization, states walk a tightrope when balancing security fears against economic incentives. Globalization promises economic gains from ­factor mobility. Economic interdependence also ­reshapes the way that states formulate g­ rand strategy. Unilateral policies can have pitfalls, resulting in short-­term and long-­term repercussions. Given ­interconnected policies, states must anticipate other states’ policy shifts. At the same time, globalization imposes formidable exigencies. The pro­cesses that facilitate f­ actor mobility also embolden and empower violent non-­state threats, making undetected entry across borders more feasible. Globalization also has an underbelly: “­every sector of the licit economy has its illicit counterpart” (Andreas 2004, 644). Naim (2005) contends that “the dark trades, driven by the same globalizing forces responsible for the surge in international commerce over the last two de­cades, now threaten the smooth functioning of the legitimate world.” What renders illicit flows particularly challenging is their clandestine nature: by evading the state’s gaze, smuggling and trafficking networks stand to weaken state authority. Clandestine actors foment anxiety over the retreat of state power (Sassen 1996; Strange 1996). Although ­t hese anx­i­eties may be exaggerated, clandestine transnational actors impede the state’s bid to exclusive territorial control.1 Eco­nom­ically open states prioritize effective border management insofar as border instability detracts from economic exchange (Simmons 2005) and stable borders require effective control over transboundary movement (Newman 2000). Far from erasing borders, globalization has served to bolster the importance of border control (Naim 2005; Newman 2006). In this security climate, it should come as no surprise that policymakers propose harder borders as a panacea against transnational threats. The new security climate also focuses on individuals as threats, even if origin

20 Chapter 1

governments are not hostile. As a consequence of this shift in focus, controlling ­human mobility is essential to maintaining security. Uncontrolled borders are vulnerable to dif­fer­ent types of transnational threats, and interlinkages among or­ga­nized crime and terrorist groups make border management all the more demanding (Shelley 2006). In this context, borders emerge as ramparts of defense against non-­state threats (Biersteker 2002). Countries retool existing border management systems to adapt to the new focus on counterterrorism (Andreas and Nadelman 2006). For instance, while pre-­inspection at airports had been implemented as a check on illegal immigration in the 1990s, the United States expanded its scope within the context of the global war on terror. Similarly, terrorism made a comeback on the Eu­ro­pean continent ­after 2001. The Madrid and London incidents animated a flurry of mea­sures labeled as antiterrorism directives. Discerning borders ward off threats and, at the same time, bridge economic flows. To accomplish this, they permit benign flows while weeding out malign flows. Discerning border policies are also po­liti­cally appealing as a means of ­handling globalized vio­lence. Terrorists rely on secrecy and surprise, which magnifies the psychological malaise that results from terrorist vio­lence. Bolstering interdependence sovereignty—­control over transborder flows—­goes some way t­ oward managing the psychological repercussions of transnational terrorism. The new security milieu has augmented the importance of border control by showing that porous borders are likely to serve as a conduit for transnational vio­lence. Keohane expands on this: “Geo­graph­i­ cal space, which has been seen as a natu­ral barrier and a locus for ­human barriers, now must be seen as a carrier as well” (2002, 32). The con­temporary foreign policy context has thus broadened the conceptualization of security to encompass transnational threats. In traditional IR terms, security is “defined in po­liti­cal military terms as the protection of the bound­aries and integrity of the state” (Doty 1998, 73). The new security paradigm blurs the strict distinction between the domestic and international insofar as a military logic is applied to coping with aty­pi­cal threats. Writing ­after 9/11, Kraska elaborated on this, noting “the traditional distinctions between military/police, war/law enforcement, and internal/external security are rapidly blurring” (2001, 501). As I outline below, the militarization of borders whereby states deploy specialized and sophisticated technology fits within this trend. From a broader theoretical standpoint, this alters how sovereignty is practiced. Traditionally, states reserved policing efforts to the



Harder Borders in a New Security Climate21

­ omestic realm and military operations to balancing against external threats d (Clausewitz et al. 2006). The new focus on asymmetric trends breaks this distinction down, with the result that borders reassert the privilege of the state (Rosiere and Jones 2012). Do security incentives make for harder borders? I answer this question by focusing on transnational terrorism to capture the con­temporary security dimension. Policy tightening can take several forms. Tougher policies commence with forging a link between dif­fer­ent types of undesirable flows. The United States accomplished this in the wake of 9/11 by rebranding law enforcement efforts as antiterrorism mea­sures, which served to stress the link between illegal migration, or­ga­nized crime, and terrorism (Andreas and Nadelman 2006; Andreas and Richard 2001). At the same time, states harness existing mea­sures by making entry requirements tougher. Higher rates of visa denials, the requirement of in-­person interviews and more extensive paperwork to apply for visas, and wider surveillance exemplify this trend. States also establish tighter controls through an extraterritorial shift of border controls whereby the state’s territorial reach expands outward (Bigo 2000). This culminates in an enlarged border zone (Bigo 2011). The recalibration of policies shifts part of the burden of control to other actors, such as other states as well as third parties such as airlines, travel agencies, and other private companies with authority to monitor travelers. The 9/11 Commission endorsed such a shift, arguing that “the US government cannot meets its ob­ eople to prevent the entry of terrorists without a ligations to the American p major effort to collaborate with other governments” (National Commission on Terrorist Attacks upon the United States 2002, 390). This shift is exemplified by prescreening checks, airline sanctions, and stricter requirements for upstream policies of control that by definition screen passengers abroad. In other words, harder borders expand the scope of who is monitored and bring more actors into the fold, thereby enlarging the scope of actors responsible for monitoring. Harder borders also manifest through vis­i­ble, on-­site policies that we would typically associate with border closure: physical barricades, cameras installed at border ports, and deployment of paramilitary personnel, or at the more aggressive extreme, minefields. Part and parcel of this pro­cess is a movement away from policing to a militarized approach to border control ­(Donaldson 2005). This transition transforms borders from sites of law enforcement and policing to sites of military operations aiming to prevent

22 Chapter 1

violent non-­state actors from obtaining access to the state’s territory ­(Lutterbeck 2004). Such high-­profile border instruments are more amenable to serving ceremonial functions. Insofar as ­these instruments affirm the state’s ability to protect, high-­profile, highly vis­i­ble policies mollify trepidation over non-­state vio­lence. On one level, high-­profile policies have a demonstrative purpose (Andreas 2009). ­These policies are rooted in po­liti­cal motives; rather than deterring access, ­these mea­sures communicate moral resolve and display authority. On another level, however, technological development and sophistication have advanced high-­profile policies. Some examples of more sophisticated technology include unmanned aerial vehicles (drones), ground-­penetrating radar (GPR) to detect underground tunnels, and new motion and heat sen­ ese instruments also permit a sors as part of a virtual fence (Elden 2013). Th smart borders approach to border control, which increasingly obviates the need for ­human personnel while also expanding the breadth and depth of the security zone. This pattern goes hand in hand with the militarization of borders insofar as the newer and more advanced technologies deployed are purchased from military suppliers and the tactics of monitoring cross over from the military realm (Jones and Johnson 2016). Not surprisingly, as a consequence of increased sophistication, border control requires an expanded bud­get. Bigo (2014) defines the Eu­ro­pean Union’s approach to border control as one of policing and monitoring, in contrast to the militarized approach of the United States. The past de­cade and a half has seen the Eu­ro­pean Union move closer to the U.S. model. As a result of this trend, Frontex, the Eu­ro­ pean Union’s border-­monitoring agency, saw its bud­get rise fifteenfold (­ Frontex 2014). To recap, highly vis­i­ble border policies can si­mul­ta­neously demonstrate military strength and fulfill symbolic roles. Rather than debordering, globalization has resulted in a rebordering, a pro­cess that rearticulates sovereign power.2 In addition, “the old model of security at discrete crossing points and dispersed monitoring of spaces in-­between has been replaced with a model that strives for ‘total awareness’ and ‘effective control’ over the entire border zone” (Jones and Johnson 2016, 194). As extraterritorial policies have been gaining currency, the state has witnessed a shift in not just how but where it exerts territorial sovereignty. At first blush, we might expect a perfect correlation and an automatic connection between the threat of terrorism and harder borders. In fact, this



Harder Borders in a New Security Climate23

expectation seemingly bears fruit ­after significant transnational terrorist attacks, with commentators predicting tighter border controls and expressing fears that the events harken back to the immediate aftermath of 9/11. ­After a terrorist attack in Ottawa in October 2014, Canada’s former deputy prime minister John Manley forecasted a clampdown on border controls, stating: “If it is in fact related to religious extremism, then I think we w ­ ill see an increased ramping up of U.S. paranoia about the border and Canada being a source of potential risk for the United States” (“Ottawa Shooting” 2014). A ­ fter the November 2015 terrorist attacks in Paris, France drew up proposals for more rigorous security checks, calling for “immediate, reinforced, systematic and coordinated controls” on the external borders of the Schengen Area. Pundits became alarmed that EU citizens would endure significantly higher logistical travel costs, including longer wait times, and systematic checks on identification documents. The United Kingdom, an EU member not part of the Schengen Area, criticized the proposals on the grounds that its citizens would bear the brunt of the burden. Moreover, states that border volatile regions encounter pressure from the international community to shore up border controls. A ­ fter the ISIS staged the Paris attacks, the United States and EU countries called for Turkey to crack down on its perforated border with Syria. Then U.S. defense secretary Ashton B. Car­ter stated about Turkey, “The single most impor­tant contribution that their geography makes necessary is the control of their own border” (Arango 2015). Thus the connection between terrorist events and harder borders is not unfounded. Nevertheless, such an automatic linkage misses the fact that policy instruments serve dif­fer­ent functions. In other words, ­border strategies differ in how they allow states to express territorial sovereignty. Consequently, it would be misleading to expect uniform policy change. Border management strategies vary according to the nature of the terrorist threat. More precisely, restrictive policies are more likely if terrorist events are salient. Vio­lence that hits closer to home and is easily observable by the public is more likely to spur policy tightening ­because such events are more likely to push policymakers to take action. In other words, events that directly imperil state interests more acutely galvanize public anx­i­eties, stoking fears over loss of control. Direct threats inflate the emblematic role that border control can play in tamping public anxiety. Previous scholarship has not shed much light on t­ hese distinct pathways insofar as it assumes an unqualified linkage between terrorism and border closure. By distinguishing

24 Chapter 1

terrorist events by venue of attack and nationality of victims, I argue that the ­ hether impact of terrorism on border management is contingent upon w threats are direct or global. We might also expect material incentives to ­counter policy tightening. September 11’s deleterious effects on economic exchange within the NAFTA area left a lasting impression. In the immediate aftermath of 9/11, the United States introduced harsher border mea­sures (Andreas 2003b). As a result, traffic across the U.S. borders with Canada and Mexico slowed to a trickle. As Andreas and Nadelman (2006) stress, this was not the first time that a crackdown by the United States halted cross-­border traffic in North Amer­ i­ca; Operation Intercept, an anti–­drug trafficking endeavor, three de­cades earlier had virtually shut down the border with Mexico. What was more significant about the post-9/11 case, however, was that it occurred in the context of economic interdependence, institutionalized and propelled through NAFTA. ­There are more recent examples where policymakers voiced alarm that terrorist attacks and their aftermath would throttle trade. The aforementioned Ottawa shooting, for example, triggered fears that border checks and red tape would stymie U.S.-­Canada trade. Likewise, ­t hese fears surfaced in the wake of the June 2015 attacks in Tunisia as well as ­after the Novem­ ere not unwarranted: ber 2015 Paris assaults (Bensemra 2016). The fears w Tunisia witnessed a significant decrease in tourism inflows during the rest of 2015 (Kim 2015). Economic interdependence and openness raise the costs of draconian border policies. Their effects on state be­hav­ior, however, are not uniform. The costs are expected to be lopsided insofar as states are asymmetrically interdependent (Gelpi and Grieco 2008). Consider the disproportionate effects of 9/11 within NAFTA, for example (Andreas 2009, 164). The economic costs ­ ere much higher than for the United States. Biof border delays for Canada w lateral trade for Canada comprises 87 ­percent of its total trade. By comparison, for the United States, the figure stands at 25 ­percent. The asymmetric commercial relationship can be expressed in terms of trade to GDP ratio. Forty ­percent of Canada’s GDP comes from (is tied to) its U.S.-­bound exports. In sharp contrast, only 2.5 ­percent of U.S. GDP comes from its exports to Canada. States strive for a balance between borders that remain open to economic exchange but yet are impregnable to penetration by undesirables. The balance also hinges on ­whether threats are diffuse or targeted. Before delving into the interplay between security and objectives, we need to spell



Harder Borders in a New Security Climate25

out why transnational terrorism should predict tighter controls and hardened borders.

Transnational Terrorism Terrorism is the “anxiety-­inspiring method of repeated violent action, employed by (semi)clandestine individual, group, or state actors, for idiosyncratic, criminal, or po­l iti­c al reasons, whereby . . . ​t he direct targets of vio­lence are not the main targets” (Schmid and Jongman 1988, 28). The ­audience of terrorist vio­lence is broader than the immediate targets of attacks. In other words, the victims are not always the intended targets of the terrorist actors but individuals who are simply at the wrong place at the wrong time (Sanchez-­Cuenca and Calle 2009).3 The latter aspect is why the repercussions of terrorist events transcend the physical damage and carnage caused by the incidents. By intimidating an audience greater than the victims of attacks, and by promising further vio­lence to come, terrorist actors aim to intimidate and force policy change from target governments (Pape 2006). In terms of military capabilities, terrorism is the strategy of the weak (Hoffman 1998). Terrorist groups, even when they solidify territorial control and draw on a global pool of recruits, cannot amass military capabilities that match ­t hose of state actors. They make use of transnational organ­izations to leverage borders to their advantage. They foment uncertainty over when and where attacks might occur through the stealth ele­ment. They explic­itly seek to catch states unaware in order to cast doubt on their ability to protect. They do so by making use of individuals to transport vio­lence across states. The degree of damage inflicted can be on par with that of state actors (Salehyan 2008b). By d ­ oing so, they highlight states’ strategic vulnerability in the face of non-­state threats. Gearson writes that the September 11 militants “utilized the long-­established terrorist approach of careful planning, ­simple tactics, and operational surprise, to effect the most stunning terrorist ‘spectacular’ in history” (2002, 7). Transnational terrorism takes advantage of the pro­cesses of globalization. It also shares the stealth ele­ment in common with other types of non-­state threats—or clandestine transnational actors (CTAs). CTAs include relatively harmless actors such as undocumented immigrants or refugees, more harmful actors such as smugglers and ­human traffickers, and more imminent

26 Chapter 1

security threats such as insurgents and terrorists. Border strategies are increasingly geared ­toward inhibiting access to such actors. As such, the theoretical connection between transnational terrorism and border control leverages insights about borders as preventive barriers against aty­pi­cal threats (Andreas 2003a; Jones and Johnson 2016). Harder borders impede the movement of CTAs across borders by raising the costs of entry and increasing the likelihood of apprehension by state agents (Hassner and Wittenberg 2015). ­These policies aim to deny such actors access to territory. Terrorism, however, is on the dangerous end of the continuum of CTAs. Or­ga­nized crime, for instance, may prey upon and detract from the ­legal economy, but terrorism is distinct in that it can cost lives, damage property and infrastructure, and even degrade the health of the country’s economy (Enders and Sandler 2006b). I contend that three interrelated features of transnational terrorism are impor­tant for understanding why transnational terrorist events prompt tighter border policies: (1) nonhierarchical spread, (2) stealth, and (3) psychological import. I flesh out each component in the sections that follow and then discuss why directed and global terrorist events have distinctive effects on the state’s responses to vio­lence.

Nonhierarchical Spread Transnational terrorism, by definition, involves crossing borders. As the preceding discussion illustrated, when a transnational terrorist event transpires, ­either the perpetrators or victims cross frontiers (Li 2005). In addition, transnational terrorist groups are nonhierarchically networked across multiple states and sometimes multiple regions (Enders and Su 2007). Groups can mobilize, recruit, and train in bases spread across several countries. Even if a terrorist group is initially limited to a specific region, it can evolve over time to develop offshoots or branches elsewhere. A noteworthy example is ISIS, which the international community initially hoped would remain confined to the Levant. ­After the Paris attacks, attention turned to the group’s external operations branch in Eu­rope (Callimachi, Rubin, and Fourquet 2016). To be sure, transnational networking does not diminish the inhibitory impact of distance (Gelpi and Avdan 2015). Although transnational terrorism calls to mind the September 11 attacks, in actuality, terrorism is a relatively short-­gun phenomenon. Terrorist groups tend to cluster in specific hot spots of high-­volume activity (Braithwaite and Li 2007). Nonetheless, even



Harder Borders in a New Security Climate27

if terrorist operatives are not mobile across borders, and even if organ­izations do not possess bases of operation in multiple countries, ­there are other ways in which terrorism can traverse borders. In general, terrorist vio­lence features spatial de­pen­dency (Sageman 2004). Even when the tangible assets b ­ ehind vio­ lence (arms, personnel, funds) do not travel across countries, the intangibles that support vio­lence—­tactics, ideas, and knowledge—­can diffuse across borders. The form that vio­lence takes can spread to other countries as homegrown groups mimic and adopt foreign groups’ tactics (Midlarsky, Crenshaw, and ­Yoshida 1980). For example, Hamas borrowed the suicide vest from the Tamil Tigers in Sri Lanka (Horo­witz 2010). The decision to go transnational—­rather than remain tied to a specific piece of territory—­can signal group resolve and capabilities. To the extent that transnational spread permits more interlinkages among groups, transnationalization can also raise group lethality. Not surprisingly, then, ISIS’s ostensible decision to go global by waging a campaign in the West compounded anxiety over the group’s growing reach. Pundits speculated as to w ­ hether the attacks on the Eu­ro­pean continent signaled a tactical shift in the ISIS’s operations, in part to counterbalance its territorial losses in Iraq and Syria (Hegghammer and Nesser 2015). Furthermore, loyalties may also traverse borders, facilitating transnational recruitment (Piazza 2006). Fi­nally, funding and weapons are also smuggled across borders, a dynamic that gains speed and currency with the emerging or­ga­n ized crime-­terrorism nexus (Dishman 2005). In sum, transnational terrorism defies and circumvents state ­ ecause groups establish bases borders in a number of ways, most starkly b across frontiers but also through transnational financing, emulation of foreign groups’ tactics, and transnational mobilization and recruitment. In contrast, the state’s power is hierarchically or­ga­nized within its own territory. Scholarship maintains that borders are more constraining on states’ armed forces than on transnational actors (Naim 2005; Salehyan 2006; Staniland 2006). Salehyan (2006) places the state’s agents at the far end of a continuum of global mobility. The moral opprobrium on territorial conquest, buttressed by the normative consensus on territorial sanctity, means that in theory at least states cannot easily move troops across neighboring territories (Zacher 2001). International borders constrain dif­fer­ent types of flows unequally: capital and then goods occupy the relatively mobile end of the spectrum, whereas the state’s security forces are at the other end. “In sum, the state is limited by its bound­aries—­the capacity to wield force. . . . ​is largely constrained by sovereign borders” (Salehyan 2006, 31). This limitation partly stems from the transaction costs involved in efficient collaboration

28 Chapter 1

among states. States confront hurdles when cooperating against aty­pi­cal threats. To begin with, states cannot agree on a common list of designated terrorist groups (Sandler, Arce, and Enders 2009). They also are not effective at coordinating antiterrorism efforts. In contrast, terrorist organ­i zations sometimes strategically pool their resources together to offset their relative military weakness (Hoffman 1998). They are effective at networking: forming alliances with each other, learning new techniques and modes of attack, and even ­going so far as to train together even in the absence of ideological commonalities. Terrorist groups excel at developing cross-­border connections and are in fact found to be, on average, more centrally connected than criminal networks (Helfstein and Solomon 2014).4 Transnational actors are able to capitalize on the disproportionate constraining effect of borders (Shelley 2006). Smugglers, for example, take advantage of price differentials across borders (Car­ter and Poast 2015), in addition to being able to relocate operations to skirt state monitoring. Transnational terrorists can similarly reap the benefits of mobility across borders by relocating and regrouping if dislodged from state territories. Terrorists’ ability to network across countries makes it difficult to detect and identify them, which in turn facilitates clandestine entry into states’ territories. To surmount the asymmetrical limitations on state power projection capabilities, states may turn to defensive mea­sures focused on preventing access to territory. Fortifying borders may emerge as a more attractive and feasible strategy when compared to risky endeavors such as air campaigns or ground incursions into neighboring territories (Staniland 2006).5 Arguably, however, technological innovations make a state’s power projection across borders more feasible.6 Drones, for example, can augment monitoring and surveillance regimes. Newer technologies overcome the normative constraint to some extent by permitting states to proj­ect power without a­ ctual boots on the ground. ­These technologies, however, are certainly not impervious to criticism.7 In addition, territorial integrity is closely linked to border fixity: as territorial conquest has become uncommon as a form of power maximization, states’ borders have become more and more fixed (Atzili 2006).

The Stealth Ele­ment The spread of transnational terrorism makes it difficult for states to mount a counterresponse, and the stealth ele­ment deepens states’ disadvantage in the



Harder Borders in a New Security Climate29

face of non-­state threats. Terrorist groups can leverage information asymmetries to gain an edge against militarily advantaged states. As a consequence, “ “an information society” such as that of the con­temporary United States would be at an informational disadvantage with re­spect to networks of individuals whose communications seem to occur largely through handwritten messages and face-­to-­face contacts” (Keohane 2002, 34). Militants can defy border control, escape law enforcement, and circumvent state surveillance. Contrary to popu­lar perception, however, clandestine territorial access does not always mean that militants slip undetected across borders. It can also transpire when terrorist groups exploit ­legal channels of territorial access. In fact, fears that terrorists would exploit the refugee regime ­were at the heart of some Eu­ro­pean states’ calls for stringency. Surreptitious entry reifies the idea of loss of control over borders and erodes sovereignty (Sassen 2006). Insofar as terrorist actors are hard to detect, conventional notions of defense do not apply to counterterrorism (Cronin 2002; Paul 2005). Additionally, terrorist events inflict psychological costs on targets b ­ ecause they carry shock value. By exploiting uncertainty over the when and where of vio­lence, terrorist actors are in effect able to gain symbolic power over state actors (Juergensmeyer 1997). This was true of several surprise attacks against democracies: Madrid in 2004, London in 2005, Paris in 2015, Ankara in 2015 and 2016, and Brussels in 2016. ­These events drove home the message that despite surveillance systems, the governments had failed to insulate citizens from ­t hese attacks. Even where socie­ties ­were on high alert and anticipated transnational terrorist vio­lence, the unpredictability of ­these events underscored the feebleness of the security establishment.8 For example, in the wake of the Brussels attack, commentators stated that while Belgium had been identified as a likely target by ISIS, and worries had been expressed over its status as a cradle of radicalization, the attacks quickly drove attention to the failures of the state’s security apparatus (Ivanovic 2016). Thus, on one level, the stealth ele­ment exacerbates anx­i­eties that even militarily power­f ul states are at the mercy of transnational militants. We imagine that militants mask their true aims at border crossings and carry out their nefarious plans once they attain territorial access. On another level, however, ­there is also uncertainty over w ­ hether l­ egal travelers are prospective recruits. The National Strategy for Combating Terrorism warned about terrorism’s global reach, that is, its capacity to have transnational and global appeal (United States 2011). Universalist non-­secular ideologies, central to the

30 Chapter 1

con­temporary wave of terrorism, permit groups to broadcast their message to broader, abstract, and often transnationally constructed communities, which they purport to represent and on whose behalf they commit vio­lence (Rapoport 2001). This lends another layer of uncertainty insofar as religion can be utilized as a transnational recruitment mechanism (Juergensmeyer 2003; Lacqueur 1999). The globalization of recruitment also makes detecting and ferreting out operatives more difficult. Foreign fighters, for example, can mask their true intentions when crossing borders. The stealth ele­ment contributes to fears over migration as a conduit for terrorism flows (Bove and Bohmelt 2015). Migration can function as an ave­ nue for transnational terrorism to the degree that it feeds the social and kinship networks that underpin radicalization and recruitment. This possibility focuses on radicalization ­after mi­grants have already gained access. A report by the Nixon Center voices ­t hese fears: “Migration and terrorism are linked; not ­because all immigrants are terrorists, but ­because all, or nearly all, terrorists in the West have been immigrants” (Leiken 2004, 6). The Hamburg Cell, a group of expatriate students that formed around a jihadi radical who had illegally immigrated to Germany, orchestrated the 9/11 attacks. Radicalization is a multistep pro­cess (Sageman 2004), whereby host-­country context can interact with active recruitment machinations by terrorist groups to produce extremism on host-­country soil. States also fear that operatives may hide among the general populace and activate sleeper cells within the host (Dishman 2005). The prospect of infiltration makes it pos­si­ble that foreign violent actors can lodge themselves within the state, in effect allowing them to repudiate borders and gain a foothold in destination states.

Psychological Impact The desire to instill and disseminate fear lies at the core of terrorist vio­ lence. Fear is what links the motivation to use vio­lence to an anticipated policy outcome (Braithwaite 2013). Fear is the pivot point of terrorist vio­ lence and, consequently, public perception is the true target of terrorist ­assaults. Terrorists use vio­lence to manipulate the expectations of an audience that expands beyond the immediate victims. They intimidate through the promise of f­ uture vio­lence to come. The public has a double role as the audience of terrorist vio­lence and the impetus for policy change (Friedland and Merari 1985). By spreading fear, terrorist actors also seek to undermine



Harder Borders in a New Security Climate31

the government’s competence in the public’s eye (Bueno de Mesquita (2005). Public fears stimulate policy change insofar as po­liti­cal leaders believe tougher policies ­w ill alleviate ­t hese fears. For public attitudes to have policy impact, leaders should also f­ actor in ­these fears. Typically, b ­ ecause demo­cratic leaders are office seeking, they are cost-­sensitive and more responsive to t­ hese fears. Management of fear is especially impor­tant given empirical patterns of terrorism: terrorism is a rare event (Mueller 2006). Risk assessment ­becomes more inaccurate in the face of high-­consequence rare events ­(Kunreuther 2002). Transnational terrorists embody this phenomenon: terrorist actors capitalize on unpredictability to create a sense of helplessness. Mueller stresses that “the costs of terrorism commonly come much more from hasty, ill-­considered, and over-­wrought reactions, or overreactions, to it than from anything the terrorists have done” (2005, 222). Precisely ­because ­t hese threats are hard to anticipate, ameliorating fear goes a long way ­toward effective counterterrorism (Friedman 2011; Khalil 2006). Scholarship from psy­chol­ogy is insightful in terms of understanding how the public responds to terrorist vio­lence. Per­sis­tent terrorism generates a range of ill effects. Moreover, ­t hese effects are enduring: terrorism not only dampens public morale but negatively (and perhaps irrevocably) alters the psychosocial fabric of demo­cratic socie­ties (Peffley, Hutchison, and Shamir 2015). We also know that terrorism evokes fear and anger and that ­these emotional responses are tied to dif­fer­ent types of policy demands (Huddy et al. 2005). Fear demands caution whereas anger demands retribution. Attitudes ­toward terrorism have implications for a range of policy outcomes. Policy change is more likely in the face of terrorism b ­ ecause the po­liti­cal milieu shifts to the right, whereby the public gravitates increasingly ­toward illiberal and authoritarian attitudes. Chronic terrorism leads to limitations on minority rights (Merolla and Zechmeister 2009). Terrorism rewrites public attitudes by sapping forbearance in socie­ties, thus posing a danger to demo­cratic governance (Peffley, Hutchison, and Shamir 2015). The lit­er­a­ture thus conveys that terrorism has direct and indirect effects on policies. More directly, vio­lence can animate specific antiterrorism mea­sures. Indirectly, it shapes public attitudes ­toward policies and thereby enlarges the scope for policymaking. The public becomes more intolerant of minorities and more willing to support hard-­line policies such as increased surveillance, enhanced interrogation tactics, and restricted civil liberties and minority rights (Peffley, Hutchison, and Shamir 2015; Piazza 2015).

32 Chapter 1

So far, scholarship has sidestepped the question of how public attitudes t­ oward vio­lence affect migration and border control. The po­liti­cal attitudes that inspire toughness in antiterrorism may do the same for migration and border control. Widespread authoritarianism and illiberalism should generate a po­liti­cal environment supportive of migration restrictions and border crackdowns. If heightened fears bring in strongman leaders and right-­w ing governments, this should also bolster policy stringency. Chronic terrorism also foments generalized feelings of insecurity, which are linked not to par­ tic­u­lar incidents or perpetrators but to beliefs that the state and society are vulnerable (Joslyn and Haider-­Markel 2007). Even when threats do not emanate from outside the state or are unrelated to terrorism, leaders can animate latent feelings of insecurity in order to push forward hard-­line agendas. The death of a border agent in Texas in November 2017 reanimated the Trump administration’s calls for the border wall (Bever, Hawkins, and ­Miroff 2017). The incident is unrelated to terrorism, but it can still be leveraged to push the wall forward ­because of extant fears of outside threats. ­There are two insights from previous research about public attitudes ­toward terrorism that I argue connect fears to policy change. The first concerns the distinction between selfish and communal fears, which mirrors the well-­k nown distinction economists draw between pocket­book concerns and macroconcerns over the national economy. Joslyn and Haider-­Markel (2007) show that sociotropic fears rather than personal fears connect more closely to policy demands. Perceptions that the community and the way of life are endangered are more power­f ul ­drivers of policy change than personal fears that one’s life is in danger. The second insight is that terrorist threats are prone to othering and, at the extreme, scapegoating t­ hose outside of the community (Piazza 2015). Vio­lence stirs resentment, creates demand for punitive mea­sures, and exacerbates ethnocentrism (Feldman and Stenner 1997). The perception that threats emanate from outside the community justifies stringency in migration and border control. Moreover, the link between terrorism and border policies relies on the perception that threats arise from outside the state. Threat perception functions through the prism of nationhood, which turns on a hard distinction between the nation as a community and foreigners as ­others. Taken together, the psychological effects of terrorist vio­lence link terrorist events to harder borders insofar as border policies serve a reassurance function. They are symbolic ­because they permit the state to demonstrate its commitment to protecting the citizenry. Public anxiety can prompt tighter



Harder Borders in a New Security Climate33

policies through a corollary mechanism: widespread anx­i­eties furnish policymakers with greater latitude in pushing forward more draconian policies. Part of this pro­cess hinges on the exploitation of fear (Altheide 2006; Mueller 2006). Fears can instigate novel policies or, alternatively, rekindle stalled policy endeavors that ­favor more rigorous border control. For example, in the case of India and Israel, border security proj­ects that had stagnated as a result of domestic opposition gained renewed steam as a direct consequence of a series of terrorist incidents: a­ fter the Second Intifada in the case of Israel and a­ fter a series of bombings following the 2008 Mumbai attacks in India’s case. Nevertheless, hypervigilance is not always about manipulating fear. Given uncertainty about the timing and location of terrorist events, it pays to be overly cautious (Friedman 2011). Policymakers would rather be overcautious than communicate optimism and be proven wrong. As Gelpi and Avdan note, “Policy makers are willing to tolerate large numbers of false-­positive predictions of a terrorist threat in order to avoid a single instance of a false-­negative prediction that results in a terrorist attack” (2015, 18). ­After all, it is far more costly for states—­and for the po­liti­cal ­careers of decision makers—to be overprepared for terrorist events that do not come to pass (false positives) than to fail to be prepared for incidents that do occur. ­After the attacks in Brussels in March 2016, the Belgian government was denounced for precisely this type of error in judgment.

Targeted and Global Threats The policy response to terrorist events is determined by how close to home the assaults occur. Incidents that directly target the state’s interests tie into tighter policies in a more straightforward manner. Direct experience can take two forms: transnational terrorist incidents conducted within the state’s borders and events that transpire abroad but victimize the state’s own nationals.9 To illustrate, on March 19, 2016, a suicide attack took place in Istanbul, killing four foreigners and injuring thirty-­six more. Among the victims ­were nationals of the United States and Israel. From the perspective of the United States and Israel, the attack occurred abroad but involved direct experience with terrorism in the sense that ­these states’ own citizens w ­ ere maimed or killed. From Turkey’s perspective, however, despite not leading to any Turkish fatalities, this incident would still be classified as direct

34 Chapter 1

experience with terrorism b ­ ecause it occurred on Turkish territory (Mickolus et al. 2007).10 To provide a contrasting example, on October 10, 2015, Turkey suffered an attack on its own soil, in the capital city of Ankara, of an unpre­ ce­dented nature. The victims ­were Turkish nationals. From Turkey’s perspective, the Ankara incident fits ­under both forms of direct experience: on its own territory and involving its own citizens. In practice, ­t hese two types of incidents overlap to a ­great extent: thus the majority of assaults that occur within the state’s borders also involve its own nationals.11 While t­ hese incidents may overlap in practice, it is pos­si­ble to differentiate between two types of ave­nues of impact: targeted incidents that I dub the territorial ­effect and compatriots’ effect.12 I propose two complementary mechanisms to support this expectation. First, targeted terrorism plays into the hands of policymakers that champion policy stringency. In effect, targeted terrorism sets in motion the pro­cess of securitization, whereby a par­tic­u­lar issue area is redefined and repackaged as an existential threat (Buzan, Ole, and Wilde 1998). Through this pro­cess, policymakers emphasize the external roots of terrorism and the dangers of permeable borders. Securitization repackages international terrorism as a grave and pressing danger, on an equal footing with military threats (Huysmans 2006). Direct threats expand the scope of policy options that are po­ liti­cally palatable. As a consequence, controversial policies that might have been a hard sell are easier to justify. To illustrate, in response to 9/11, liberal democracies ­adopted tough legislation dealing with immigrants and foreigners and expanded the rights of the executive to survey and assem­ble information (Epifanio 2011). Some of ­t hese mea­sures might have been po­liti­cally unpop­u­lar, without the threat of transnational terrorism, b ­ ecause they cut into the procedural and privacy rights of not just foreigners but also citizens. However, the sense of urgency generated by attacks on ­t hese countries’ own territories—9/11, 4/11 (Madrid), and 7/7 (London)—­downplayed the influence of liberal reservations. Second, targeted events have a more pronounced impact on public fears and anx­i­eties. To the extent that border control aims to quell t­ hese fears, we should expect incidents that trigger alarm to connect more closely to harder borders. Psychologists connect fear to a demand for caution and defense (Huddy et al. 2005). Tighter controls align with a defensive perspective and borders are viewed as a protective shield against external threats (Staniland 2006). Researchers also document a significant shift in public attitudes as a



Harder Borders in a New Security Climate35

consequence of terrorism, such that p ­ eople are more willing to tolerate painful mea­sures, including restrictions on civil liberties (Davis and Silver 2004). Accordingly, the public ­w ill be more likely to welcome and even embrace draconian policies that run ­counter to liberal norms. Harder borders depart from liberal princi­ples of freedom of movement. Absent direct attacks, the public may not support them. An ancillary line of research finds that transnational terrorism realigns the po­liti­cal environment by making the public more likely to elect right-­wing, hawkish po­liti­cal leaders (Huddy et al. 2005; Lahav 2004). It is plausible that hard-­line leaders are more likely to trumpet tougher border controls.

Territorial Effect

Transnational terrorist incidents that transpire on the state’s own territory are likely to heighten negative attitudes ­toward terrorist vio­lence more than terrorist events that transpire abroad. Physically proximate events are more salient. An extensive line of research shows that ­those proximate to areas suffering from prolific levels of terrorism exhibit higher levels of anxiety and fear and that ­t hese emotions endure (Allouche and Lind 2010). Analogously, proximate vio­lence is more vivid in memory (Lowenstein et al. 2001). Empirical studies of public attitudes show that Americans exhibited symptoms akin to post-­traumatic stress disorder (PTSD) in response to 9/11 and, further, that ­t hese symptoms ­were more acute among t­ hose who lived in New York’s metropolitan areas (Schuster et al. 2001). Vio­lence that takes place on ­ ill also be more jarring and therefore enduring in collective one’s own soil w memory. Numerous studies, conducted in countries that bore witness to major terrorist events—­Spain, the United Kingdom, and the United States—­ corroborate this insight (Allouche and Lind 2010; Braithwaite 2013; Romanov, Zussman, and Zussman 2010).

Compatriots’ Effect

Incidents that victimize the state’s citizens are a form of direct experience with terrorism. I contend that even when assaults occur in other countries, the involvement of the state’s own nationals ­will generate a stronger policy

36 Chapter 1

response. Such events are simply publicly more vis­i­ble; the death of a country’s citizens invites longer and more expansive media coverage. ­People are more likely to be cognizant of t­ hese events. Such attacks may be physically distant but are personally proximate. Personal proximity centers on shared communal traits, in this instance defined by citizenship. ­People come to believe they may be next in line. Such events also pose a danger to citizens’ broader community and way of life (Huddy et  al. 2002). Assaults that ­v ictimize the state’s own citizens are likely to elicit a stronger outcry ­because the public can readily empathize with fellow citizens. Parallel to attacks on the country’s soil, ­these events create room for tougher policies and play into the hands of hard-­line policymakers who espouse a tougher approach to border control.

Trading Security for Economic Gain Smart borders demand that states pursue selective policies instead of a blanket clampdown. Border closure targets specific source states and takes into account the past history of attacks. In contrast to global threats that are diffuse, incidents that are traceable to source countries permit selective border closure. A selective strategy is more appealing to eco­nom­ically open states ­ holesale border closure. Naim (2005) argues that selecthat cannot afford w tive border policies are necessary for coping with the exigencies of globalization. He maintains that such policies are necessary for “a government that is mandated to control an increasing number of cross-­border activities, on all borders” (Naim 2006). Precisely ­because states are limited in their ability to identify and track down clandestine transnational actors, we expect them to craft policies that minimize territorial breach by potential threats. However, smart borders do not simply screen undesirables. They also sustain borders open to benign flows. Therefore, permeable borders necessitate that, to some degree, states must trade security for economic gain. Economic interdependence scholarship helps illustrate how states balance economic goals with security objectives. A rich body of knowledge in international relations maintains that trade and capital ties shape states’ conflict be­hav­ior (Doyle 1997; Mansfield and Peve­house 2000; Mansfield and Pollins 2003; Polachek 1980; Viner 1951). Extrapolating from this, I argue that commercial ties influence states’ security-­seeking be­hav­ior vis-­à-­vis transnational threats. Economic and security interests intersect when it comes to control-



Harder Borders in a New Security Climate37

ling borders. The argument builds on the work of liberal scholars who argue that trading states are more reluctant to use militarized force (Domke 1988; Rosecrance 1996; Rosecrance 1986). Mirroring this contention is the idea that commerce and territorial conquest accomplish the same objectives, so that material incentives eclipse security interests in state ­g rand strategy. A more ­nuanced argument is rooted in the logic of opportunity costs whereby bilateral economic ties are central to the pacifying effect of trade (Oneal and Russett 1997; Oneal and Russett 1999; Russett and Oneal 2001). Specifically, states shy away from engaging in militarized disputes against commercial partners. Opportunity costs are an impor­tant reason why states are reluctant to take up arms against economic partners. States anticipate backlash from partners; the prospect of economic loss from disruption or diminution of trade in turn restrains the use of force. Neoclassical economics would tell us that economic openness predicts open borders. Empirical patterns, however, fail to sustain this expectation. To stress an earlier point, migration scholars argue that trading states continue to pursue relatively closed migration and border policies (Cornelius et al. 2004; Hollifield 2000). The neoclassical argument may miss the mark insofar as it ties economic openness to liberal migration policies without heeding how bilateral ties affect states’ policies. In order to overcome this theoretical shortcoming, I contend that general openness to economic flows may not necessarily entail open borders; instead, economic interdependence ­ ill predict more liberal policies ­toward the with a commercial partner w partner. Specifically, asymmetric dependence on commercial partners ­w ill inhibit draconian policies against their citizens. Two distinct lines of argumentation support this proposition. Hardened borders and restrictive policies are hostile signals. States view visa restrictions, for example, as punitive sanctions, and the imposition of visa requirements may trigger retaliation from dyadic counter­parts. A more extreme example is border barricades: ­these impediments aggravate neighbors even in the absence of territorial disputes (Donaldson 2005). To the extent that neighbor states perceive barriers to be adversarial signals, tighter policies bear opportunity costs. If ­these costs restrain states’ conflict involvement, we can expect similar restraints to act upon states’ border-­control policies. A somewhat less well-­known strain of economic interdependence scholar­ ship is grounded in so­cio­log­i­cal liberalism (Viner 1951). Scholars writing in this tradition argue that economic exchange hinges on communication and contact between states, which in turn enhance trust (Fordham and

38 Chapter 1

Kleinberg 2010). Through commerce, states gain knowledge about each other’s customs and practices. Harder borders are antithetical to liberal precepts and collide with the norms of a ­free market society (Flynn 2003). As such, eco­nom­ically integrated states should be guided by shared liberal ­tenets. At the same time, greater trust entails that states downplay security fears from economic partners. The state-­level arguments can only take us so far. A ­ fter all, t­ here are highly interdependent states, such as U.S.-­China and U.S.-­Mexico, that do not have liberal migration and border policies with regard to each other. In other ­ ecause migration control is not high words, state-­level arguments are limited b politics and is thus on a dif­fer­ent footing compared to the decision to use mil­ ere is, in addition, a societal layer to how interdependence itarized force. Th connects with more open borders. Economic ties create vested interests in maintaining open borders. Tighter policies hinder personal contact and interaction, which are crucial to international economic exchange. Stakeholders are thus likely to oppose policy tightening due to the costs on face-­to-­face contact with business partners. Fervent proponents of trade and capital liberalization have historically trumpeted open migration policies (Hollifield 1992; Hollifield and Zuk 1998). Insofar as trade and migration are complementary, we expect pro-­trade lobbies to champion looser migration policies (Rudolph 2008).

Looking Ahead To recapitulate, border control fulfills two functions. The first one is demonstrative and premised on showing the state’s territorial authority. The second one centers on effectiveness and necessitates screening out and denying access to threats. Border instruments are differentiated according to how well they can perform ­t hese functions. The theory of border control makes three overarching claims with empirical implications. First, fears over security lead to policy restrictiveness. The book concentrates on terrorism as a crucial manifestation of security concerns in the con­temporary policy climate. The second claim adds nuance to the first by arguing that threat perception varies according to ­whether vio­lence is directed ­toward state interests. The third claim is grounded in economic interdependence theory and suggests that trade and capital ties make for more open borders. They do so, directly, through the opportunity costs generated by harder borders. They also do so



Harder Borders in a New Security Climate39

by modulating the impact of concerns over terrorism. While the theory developed in this chapter applies broadly to border control, the empirical chapters develop hypotheses on specific instruments of control and test the implications of the theory separately on ­t hese instruments. Looking ahead, Chapter 2 formulates hypotheses on visa restrictions. Chapter 3 focuses on visa requirements and visa rejection rates. Chapter 4 shifts the focus to border barriers in the twentieth and twenty-­first centuries.

CHAPTER 2

Terrorism, Trade, and Visa Restrictions

Visas and Border Control In the previous chapter, I argued that border closure w ­ ill be more likely in the face of transnational terrorism, especially when attacks are directed against a state’s territory and citizens. Economic ties w ­ ill offset the impact of security concerns and make for open borders. This chapter ­w ill explore the observable implications of the theoretical framework for visa restrictions. Visa restrictions comprise an impor­tant, albeit understudied, component of migration and border control. A visa is defined as a “document issued in the country of origin (or residence) of the individual by the authorities of the state to which he or she wishes to go” (Guild 2009, 118). If a state has imposed a visa requirement against an origin state, the citizens of the origin state need to apply for a visa to legally gain entry. Visas thus constitute an upstream form of control, in the sense that they regulate and monitor mi­grants before they have gained entrance to the destination state. Travelers apply for permission to legally enter destination states’ territories at consulates and embassies abroad. The upstream aspect of visa controls permits states to push border controls outward, beyond physical borders, both into cyberspace and into foreign space (Bigo 2011). At the same time, governments enlist third parties, for example, obliging private agents such as transport companies, in monitoring travel. ­Legal entry often requires varying degrees of documentation, depending on the host, origin state, and visa category (Hobolth 2013). Importantly, the visa does not guarantee admittance but merely allows travelers to legally ask for permission to enter destination states’ territories. This grants border officials some discretion when monitoring border crossings. However, sometimes states also utilize visa controls at border ports as an on-­site mechanism of control over border crossings. Sometimes referred to



Visa Restrictions41

as the “sticker visa,” this policy instrument does not require that travelers obtain ­legal permission to enter destination state territories beforehand. Hence, the visa at the border does not embody the screening goal that the more common upstream visa requirement serves (Neumayer 2006). Instead, the visa-­at-­the-­border generates revenue by requiring a nominal fee from border crossings, similar to the toll on an interstate highway in the United States. As such, the analyses herein focus on the upstream visa instrument. The Universal Declaration of ­Human Rights stipulates a right of exit and entry to one’s own country (Article 13) but not a right to enter foreign space. States have the prerogative to arbitrate who gains access to their territories. Visas and border controls are impor­tant tools for states to exercise control. They are also historically intertwined with the emergence of the modern territorial state. With the development of standing armies, the state took over the legitimate means of coercion. Concomitantly and similarly, the modern state has monopolized the legitimate means of mobility, restricting entry and exit across international frontiers (Torpey 2000a). Identity cards ­ ere essential for documenting memand codes, and eventually the passport, w bership and developing mechanisms to distinguish among them for administrative purposes. Visa restrictions are, relatively speaking, a modern invention. Together with the passport, visas and biometric identity documents uphold the mobility regime (Huysmans 2006). Visas monitor short-­term mobility. In the empirical analy­sis that follows, we ­w ill address visa controls as they pertain to short-­term territorial access. The next chapter also examines visa applications for dif­fer­ent categories of short-­term migration, including work, study, business, and tourism. Importantly, however, visa policies impact other dimensions of migration control. For example, states tighten visa policies in order to ward off prospective asylum seekers (Finotelli and Sciortino 2013). For precisely this reason, migration scholars consider the visa regime an impor­tant component of the state’s overall asylum and refugee-­control regime (Czaika and Hobolth 2016). Visa restrictions thus dissuade prospective asylum seekers and divert them to states with softer border-­control policies. Visas also are essential for reducing illegal migration. Contrary to popu­ lar belief about mi­grants slipping across borders unnoticed, visa violations are one of the primary ways in which ­people gain illegal access to advanced democracies (Jordan and Düvell 2002; Torpey 1998). According to the Department of Homeland Security, for example, in 2015, of the 10.9 million illegal mi­grants in the United States, 40 ­percent gained access via visa transgression

42 Chapter 2

(Attanasio 2016). Neumayer (2006) contends that visa overstay is one of the primary c­ auses of irregular migration into Eu­rope, where clandestine border crossings have been traditionally less frequent than along the U.S.-­Mexico border. In general, stricter visa regimes compel mi­grants to seek irregular ways of gaining access to and staying in destination states. Czaika and Hobolth (2016) find a 4–7 ­percent jump in the irregular migration rate into the Eu­ro­ pean Union following a 10 ­percent spike in the visa rejection rate. States use visas to regulate entry into destination state territories.1 Visa restrictions perform a screening function, acting as “the first line of defense” against the entry of undesirables (Torpey 1998, 252), which include prospective undocumented mi­grants as well as security risks. States design visa regimes to lower the ex ante probability of mi­grants transgressing the terms of the visa. To do this, they must be able to weed out individuals most likely to overstay and thereby transition into illegal status. For this reason, for instance, origin states that generate a high volume of undocumented migration typically confront a higher likelihood of visa requirements from advanced democracies (Neumayer 2006). Visas also have a preselection function. Czaika and Haas (2013) argue that, in general, migration control aims to distinguish between desirable and unwelcome mi­grants. Through selective controls, states target the composition of flows, usually in terms of skill but sometimes in terms of ethnicity and religion. Similarly, visa controls allow governments to encourage the mobility of some while inhibiting and deterring the mobility of o ­ thers. Boehmer and Peña (2012) conceptualize border openness as a continuum, reflected in the amount of documentation the destination state requires for ­legal entry. Fully open borders allow passage with minimal documentation; the Schengen zone has heretofore permitted citizens of Schengen member countries to travel without a passport. In other instances, states may have a separate category of origin states from which they demand only an identity card in lieu of a passport to gain entry. The requirement of a passport to gain admittance indicates a stricter approach. To illustrate, in 2009, the United States introduced the requirement that individuals carry a passport to enter from Canada, a development that signified the tightening of border controls in NAFTA. Visa requirements are a more restrictive requirement, representing a step t­ oward border closure. Not surprisingly, t­ here are no cases where states impose a visa for travel without a passport. The global visa regime has become more permissive over time. However, this liberalization has proceeded in an uneven fashion (Mau et  al. 2015; Neumayer 2006). While the citizens of advanced Western democracies gained



Visa Restrictions43

v­ isa-­free travel privileges to a greater number of countries, the rest of the world’s citizens encountered new restrictions. In other words, visa policies are not necessary reciprocal. Where reciprocity applies, it is restricted to a smaller set of advanced democracies. What we do not yet know is ­whether and how transnational terrorism contributed to this bifurcation in the global visa system.

Empirical Implications of the Theory This section develops a set of testable propositions concerning bilateral visa restrictions that serve as the basis of the empirical analy­sis. The first set of hypotheses ties transnational terrorism to visa policies. U ­ nder what conditions does terrorism impel stricter visa policies? First, we can envisage a global effect. By design, visa requirements are a tool for defending against security risks. Nevertheless, ­there are several ways that visa policies can be used more or less discriminately. We expect states to capitalize on the preselection, screening, and deterrent roles of visa controls to regulate territorial access. Prospective mi­g rants who are exempt from visa requirements are deemed nonthreatening and ­those facing restrictions undergo additional scrutiny and background checks (Neumayer 2006). The previous chapter argued that states confront an endemic type of uncertainty with re­spect to transnational terrorist groups. Asymmetric informational disadvantages make it more difficult for states to detect and ferret out transnational militants. Through clandestine access, terrorist actors can flout states’ surveillance technologies. Terrorist groups make up in secrecy and surprise what they lack in military power. By organ­izing across states, they place states at an operational disadvantage. Hence, uncertainty over the where and when of terrorist vio­lence reinforces the importance of visa controls. If visa policies are the first line of defense against potential threats, as Torpey (1998) surmises, then states should capitalize on visa controls to address this type of uncertainty.

Hypothesis 1: Global Effect

States w ­ ill enact restrictive visa policies against origin states whose nationals have perpetrated incidents of transnational terrorism. One pos­si­ble response is to restrict mobility from terror-­exporting origin states. If we conceptualize transnational terrorism as a flow of vio­lence,

44 Chapter 2

then states whose nationals have been involved in a high volume of incidents worldwide are significant exporters of terrorism (Muller 2010). I label this policy response the global effect of transnational terrorism. Namely, destination countries are attentive to ­whether the nationals of an origin state have frequently executed transnational terrorist events. This expectation is consonant with the notion that in the current security climate, states respond to individuals as threats (Salehyan 2008a). States may place restrictions on certain countries even if ­t here is no animosity ­toward their home governments. From a slightly dif­fer­ent perspective, the rogue-­state framework that gained ­ ecause of their purcurrency post-9/11 labels specific countries as pariahs b ported support of terrorism (Caprioli and Trumbore 2005). The international community may cast blame upon terror-­rich states for turning a blind eye to their nationals’ involvement in worldwide terror.2 The contagion of terrorist vio­lence across borders magnifies the global effect on policies. Transnational terrorist events may spill over into neighboring states even if the grievances are homegrown (Braithwaite and Li 2007). Alternatively, terrorist groups strategically target softer targets (Enders and Sandler 2006). Hence, proximate states’ more stringent policies may redirect terrorist vio­lence ­toward states with more permissive conditions for vio­lence. Thus events in other states inspire heightened perceived probability of f­ uture vio­lence. Even if the state in question is not targeted in attacks, vio­lence in other countries increases the perceived security risks of ­human mobility. In sum, we expect states to tighten visa policies in response to attacks elsewhere.

Hypothesis 2: Targeted Effect

States whose citizens or territory have been harmed in incidents of terrorism ­will enact restrictive visa policies against origin states whose nationals ­were associated with ­these incidents. The global effect is selective in the sense that states tailor policies to specific countries, imposing restrictions on terror-­rich origin states. That is, it does not imply a w ­ holesale response whereby attacks abroad drive destination states to enact more stringent policies against all sending states. However, it is somewhat less discriminate in that it means stricter policies against origin states, regardless of ­whether the destination state has been directly harmed by t­ hese attacks. We can imagine a more selective type of policy whereby states respond to attacks by origin-­country nationals only if their



Visa Restrictions45

national security interests ­were involved in ­these attacks. Incidents that harm the state’s citizens, or take place on the territory of the state, directly threaten its national security. I refer to this as the targeted effect of terrorism. Targeted events also more directly activate the fear-­management aspect of border controls. Friedman (2011) argues that responding to terrorism is mostly about managing perceived vulnerability to attacks. Targeted assaults generate a sense of vulnerability, propelling states to calibrate policies according to the pos­si­ble harm rather than to precise risk. However, Andreas (2009) contends that policy stringency is, to an extent, an outgrowth of public demand. Regardless of the objective efficacy of policies, restricting territorial access to outsiders also fulfills domestic demand. Policymakers also utilize draconian mea­sures to symbolically assert their commitment and ability to insulate the citizenry from external threats. The symbolic aspect is muted when it comes to visa controls b ­ ecause the public is not always cognizant of visa restrictions. The government may make much ado about imposing visas on terror-­exporting states, but even then the visa requirement is physically not salient. That said, insofar as attacks on the country’s soil or involving its citizens elevate threat perception, they inspire the public to demand that the government take action. Furthermore, targeted attacks compound the securitization of ­human mobility, which pivots on countries reframing migration as a threat to physical integrity (Lavenex 2001). Terrorist events on the state’s own soil directly allow securitization of short-­term mobility. Targeted attacks muster public backing of harder policies by creating widespread fears of being victimized. In addition, they agitate the public into demanding tighter policies. In effect, the public push allows governments to put aside the objective efficacy of policies. Additionally, terrorist events tend to cluster together spatially and temporally (Braithwaite and Li 2007; Gelpi and Avdan 2015). Put simply, unlike lightning, terrorist events do strike the same locale twice. Hence, it is logical to expect states that have suffered vio­lence at the hands of origin-­country citizens to selectively curb ­human mobility from ­these states.

Hypothesis 3: Economic Interdependence

Eco­nom­ically interdependent states ­will be less likely to pursue restrictive visa policies with re­spect to their economic partners’ citizens.

46 Chapter 2

At the heart of the theoretical framework is the argument that economic ties make for more liberal policies. They do so through direct and indirect effects; that is, by affecting migration policies and by conditioning the impact of terrorism. Insofar as stiffer visa policies degrade economic exchange, we’d expect opportunity costs to ­matter in states’ decision making. That is, anticipating revenue from economic exchange to shrink, states ­w ill be less willing to implement stringent policies. Reduced travel is a detriment to economic exchange to the extent that trade and foreign direct investment rest on face-­to-­face contact. The same characteristics of visa controls that augment security can inhibit travel. Visa restrictions raise the costs of travel as a result of the wait time, paperwork, fees, and uncertainty involved in the application pro­cess. Not surprisingly, visa requirements may discourage prospective travelers. In addition, tougher visa policies significantly dampen tourism. Neumayer (2010) demonstrates that, on average, visa controls decrease travel between pairs of states by 52–63 ­percent. We may contend that tourism-­dependent destination countries are more likely than other states to hesitate to impose restrictions.3 However, this misses the strong positive correlation between travel and commerce. As O’Bryne notes, “Freedom of travel is freedom to trade” (2001, 409). Visa re­ ecause while goods quirements function akin to non-­tariff barriers to trade b and capital can circulate, the producers of goods and ­owners of capital can­ ehind t­ hese words if we consider not. We can further unpack the wisdom b that even in the electronic age, the physical presence of investors encourages the establishment and preservation of business interlinkages. Neumayer (2011) reports the reductive impact of visa restrictions on the flow of goods and capital. He argues that primarily by hampering personal contact across borders, visa restrictions significantly lower bilateral trade and foreign direct investment (FDI). He finds that visa restrictions result in a drop of ­21–32 ­percent in bilateral trade and a moderately higher drop of 33–38 ­percent in bilateral FDI. Neumayer’s (2011) findings point to the direct negative correlation between restrictive visa policies and economic transactions. Stricter visa policies can erode the dyad’s economic relationship through an alternative mechanism: to the degree that draconian policies are viewed by the commercial partner as an antagonistic signal, we might expect a backlash. A ­ fter all, restrictions on mobility are antithetical to liberal tenets and collide with shared norms between trade partners (Flynn 2003). Consequently, restrictions might anger economic partners. To illustrate, in response to the Eu­ro­



Visa Restrictions47

pean Union’s controversial proposal to impose visa restrictions on the citizens of the United States and Canada, critics ­were quick to voice concern over the breakdown of transatlantic relations (Kanter 2016). Additionally, the commercial partner may retaliate by enacting tougher legislation, further stifling economic revenue. Neumayer’s (2010) empirical analy­sis validates this intuition, showing that reciprocal visa restrictions reduce capital flows between states by 6–12 ­percent more than unilateral visa requirements. Reprisals can also manifest as weaker economic ties, e­ ither ­because investors become reticent to continue ­doing business in the partner state or ­because the state revokes economic privileges. Key domestic actors in the origin state may also feel slighted by stiff policies. By inhibiting face-­to-­face communication and contact, restrictive policies ­w ill undercut mutual affinity. Turkey’s response to severe visa requirements by Schengen states is a case in point (Kirișçi 2007).4 Pro-­trade lobbies and business in Turkey have time and again castigated the Eu­ro­pean Union for tough visa legislation that contravenes the spirit (and, to some extent, the ­legal framework) of the Eu­ro­ pean Customs Union. ­These concerns prompt commercial lobbies at home to champion liberal visa policies t­oward economic partners. In sum, states ­w ill consider the opportunity costs of reduction in trade and capital investment when devising visa policies. Policymakers ­will also be attuned to vested interests favoring liberal policies.

Hypothesis 4: Conditioning Effect

Economic ties w ­ ill decrease the impact of transnational terrorism on visa policies. Economic interlinkages also impose indirect effects by modulating the impact of security imperatives. Material incentives carry more weight than security imperatives in trading states’ g­ rand strategy (Rosecrance 1986; ­Rudolph 2003). Societal actors that stand to lose from tighter policies may boost the salience of material concerns in the state’s calculus. For example, trade lobbies and multinational corporations have admonished the U.S. government for tough policies that stymie and slow down trade (Yu 2010). Trade groups in par­tic­u­lar raise the concern that visa restrictions risk rechanneling business to economic competitors of the United States, such as Brazil, China, and India. Even when the visa is granted, it might have a cooling-­down effect and divert foreign business abroad, thus weakening a country’s ­competitive

48 Chapter 2

edge over its economic rivals. Domestic co­ali­tions with internationalist preferences may forge cross-­national bonds, facilitating economic interdependence and prosperity (Solingen 1998). Th ­ ese ties in turn significantly shape foreign policy attitudes, both among po­liti­cal elites and at the public level (Fordham and Kleinberg 2009). Economic exchange inculcates mutual trust, which in turn should temper fears associated with economic partners’ citizens. Hence, we expect economic interests to downplay the impact of transnational terrorism: the positive impact of transnational terrorism on policy stringency declines with the strength of economic ties.

Mea­sur­ing the Concepts The hypotheses pertain to visa policies in general. This chapter focuses on visa restrictions; the subsequent chapter proceeds to visa rejection rates and documentation requirements. The dependent variable (DV) employed in the analyses that follow is coded 1 if the destination state has a visa requirement in place for the citizens of the origin state and 0 other­w ise. The data cover 189 member states of the United Nations and 18 nonmember po­liti­cal territories.5 This yields information on bilateral visa restrictions for 36,300 directed dyads. Directed-­dyad design allows each state to appear once as destination (recipient) and once as origin (sending) state. That is, the data differentiate between a visa in force by state A against state B and vice versa. The directed-­dyad design is appropriate for this study b ­ ecause visa reciprocity is not guaranteed. To illustrate, 32 ­percent of states do not reciprocate visa waivers; reciprocity is more common among Western democracies and the Organisation for Economic Co-­operation and Development (OECD) member states. Put another way, 68 ­percent of states do reciprocate policies, but this figure also includes reciprocating by imposing visa restrictions (close to 25 ­percent). The data on visa restrictions are cross-­sectional and allow me to interrogate variation across pairs of states (dyads). Ideally we would track visa ­restrictions over time.6 However, as Neumayer (2006, 2010, 2011) has noted, assembling such a data set would be a prohibitively costly and time-­consuming endeavor. I utilize Mau et al.’s (2015) data set on visa waivers for 2010, coded from the International Aviation Travel Association’s (IATA) Travel Information Manual (TIM), which provides authoritative information on states’ visa



Visa Restrictions49

policies (IATA 2010). The main analy­sis focuses on visa restrictions for 2010. Mau and coauthors’ coding registers 1 if the destination state extends visa-­ waiver privileges to the origin state and 0 if visas are not in place. A visa waiver is defined as visa freedom if individuals can travel to the destination state without an application procedure before departure. Visa waivers usually allow a stay of ninety days, which is also standard for a short-­term visa. I reversed this variable so that 1 represents a visa restriction by state A for citizens of state B and 0 other­wise. Mau and coauthors’ data set excludes small island states (Tuvalu), territories that lacked international recognition of statehood as of 2010 (Kosovo), and very small states (Andorra). I filled in ­these cases from Proj­ect Visa, a website that provides information on visa-­f ree travel privileges between countries.7 Neumayer (2006) used the 2004 edition of TIM to code visa restrictions for 2004 to create a more comprehensive data set that also covers small ­island states and nonin­de­pen­dent territories. I used the data in Avdan 2014a, ­ ere, which examines the impact of terrorism and trade on visa restrictions. H I refer to ­these data to compare visa restrictions across ­these two time points. Although visa policies evolve slowly, the comparison of 2010 with 2004 does show change over time, mainly t­ oward liberalization. For example, the 2010 data show 783 new directed-­dyad pairs that have visa-­free travel; given the total number of directed-­dyad pairs (36,300), this figure reflects 2.15 ­percent of the sample of countries that have liberalized their visa policies. Nevertheless, to ensure that the discrepancy was not an artifact of dif­fer­ent coding procedures by the researchers, I consulted alternate sources on incongruent cases. Specifically, where visa policies as of 2010 conflicted with ­t hose of 2004, I cross-­checked the cases with Proj­ect Visa. To illustrate, in 2005 and 2010, Turkey lifted visa requirements for a number of countries in the ­Middle East, including Jordan and Lebanon; the variables for 2004 and 2010 would thus differ, taking the values of 1 and 0, respectively. The updated 2010 data are thus fairly recent and reflect ­t hese changes in countries’ visa policies.

In­de­pen­dent Variables

The first set of hypotheses concentrates on the impact of transnational terrorism. The data for transnational terrorism are sourced from International Terrorism: Attributes of Terrorist Events (ITERATE). The ITERATE data

50 Chapter 2

cover transnational terrorist events for 1968–2011 and provide information on event characteristics (Mickolus et al. 2012). Unlike other comprehensive databases on terrorism, such as the Global Terrorism Database (GTD), ITERATE focuses exclusively on transnational terrorism. A key advantage of ITERATE is that data harness information from international as well as domestic media outlets, which is particularly impor­tant when accounting for terrorist vio­lence in countries with state-­controlled media. According to ITERATE, terrorism is defined as “the use, or threat of use, of anxiety-­ inducing, extra-­normal vio­lence for po­liti­cal purposes by any individual or group, ­whether acting for or in opposition to established governmental authority, when such action is intended to influence the attitudes and be­hav­ ior of a target group wider than the immediate victims” (ibid., 2). By definition, the database excludes crimes or vio­lence committed by state agents. ITERATE also excludes vio­lence that occurs during civil or interstate war. For instance, vio­lence conducted by guerilla groups would not be included ­u nless it is aimed at civilians. Transnational terrorist events are ­t hose that cross borders through the nationality or foreign ties of perpetrators, institutional or ­human victims, location, the dynamics of resolution, or repercussions. Hence, domestic events are excluded from the overall count of terrorist events. To illustrate, in July 2011, Anders Breivik, a Norwegian citizen, committed a terrorist assault ­ ecause it in Oslo. The attack does not make it into the ITERATE database b does not qualify as a transnational incident. The victims of the attack and the perpetrator ­were all Norwegian citizens, and the incident took place on Norwegian soil. In contrast, consider the Boston marathon bombing of April 2013. While one of the perpetrators was a naturalized American citizen, the other was a foreign citizen. In addition, the victims included both ­ eople who American citizens and citizens of other countries; one of the three p ­were killed was a Chinese national. The incident thus qualifies as transnational and would make it into ITERATE by virtue of the nationalities of the perpetrators and victims.8 Incidents carried out by the state’s own citizens—­ so-­called homegrown events—­can also be transnational if they involve foreign victims. The London tube bombing in July 2005 provides an illustration. Although the event was considered homegrown ­because British ­nationals staged the attacks, citizens of other countries ­were also victimized. In short, ITERATE includes incidents in which e­ ither the victims or perpetrators of attacks have crossed borders and incidents in which both victims and perpetrators have crossed borders.



Visa Restrictions51

To assess the first set of hypotheses, I create separate global and targeted counts of transnational terrorism. The global count aggregates the total number of incidents committed by the nationals of the origin country anywhere in the world. The mea­sures sum incidents from 2000 u ­ ntil 2009. This operationalization allows for a check on potential reverse causality: the dependent variable registers visas implemented in 2010 and the sums are therefore lagged by one year. To capture targeted terrorism, I utilize two mea­sures. The first rec­ords incidents by origin nationals against destination nationals. The second rec­ords incidents by origin nationals on destination territory. ­These two indicators also correspond to the nationality-­and location-­based indicators widely used by terrorism scholars to create directed-­dyadic mea­sures of transnational terrorism (Young and Findley 2011). The San Bernardino attack in December 2015 would be reported as a targeted incident, executed by a Pakistani national, and transpiring on U.S. soil and involving U.S. victims. The March 2015 bombing in Istanbul, which killed and injured American citizens, also fits the targeted operationalization: although occurring abroad, it would be counted in the nationality-­based targeted mea­sure of transnational terrorism, with the United States as the target state. Similarly, the June 2015 ­ ecause it victimized British attack in Tunisia has Britain as the target state b citizens. One of the assailants was a Tunisian student. Therefore, if the targeted count ­were extended to include 2015, this incident would count as a directed-­dyadic incident by Tunisia against the United Kingdom. To keep the operationalization of variables consistent, the targeted counts aggregate incidents from 2000 ­until 2009. In sensitivity analyses, I reevaluate models using mea­sures that include the years 1990–2009. A further coding decision is worth mentioning. ITERATE rec­ords up to three nationalities for perpetrators and victims. My mea­sures account for all three nationalities, but the results are not altered if only the first nationality is accounted for.9 Fi­nally, targets of attacks may not necessarily be the victims of attacks; ­ ere. For instance, it might be impossible to tell who the intended victims w Americans and Israelis ­were victimized in the Istanbul bombing in March  2015, but we do not know if the ISIS explic­itly sought to target ­Americans and Israelis in this case. This issue plagues the wider spectrum of empirical studies of transnational terrorism (Young and Findley 2011). However, for my purposes, what m ­ atters is if the state’s interests (citizens or territory) are directly involved, regardless of the ­actual aims of terrorist operatives. Hence, ­there is no reason to expect the issue to systematically affect inferences from multivariate analy­sis.

52 Chapter 2

­Table 2. Summary Statistics for Transnational Terrorism Indicators Variable

Mean

Std. Dev.

Min.

Max.

14.71 4.38 3.84 1.68

33.29 12.03 9.34 5.18

0 0 0 0

257 92 69 38

0.10

1.21

0

78

0.03

0.58

0

44

0.09

2.31

0

220

0.02

0.67

0

63

0.03

0.50

0

45

0.02

0.30

0

26

0.02

0.54

0

41

0.01

0.31

0

33

Global Terrorism Global terrorism, post-1990 sum Global terrorism, post-2000 sum Fatal incidents, post-1990 sum Fatal incidents, post-2000 sum Targeted Terrorism Targeted terrorism, post-1990 sum (by nationality of victims) Targeted terrorism, post-2000 sum (by nationality of victims) Targeted terrorism, post-1990 sum (by event venue) Targeted terrorism, post-2000 sum (by event venue) Fatal incidents, post-1990 sum (by nationality of victims) Fatal incidents, post-2000 sum (by nationality of victims) Fatal incidents, post-1990 sum (by event venue) Fatal incidents, post-2000 sum (by event venue)

Note: The summary statistics include the West Bank and Gaza (Palestine), which is not included in the statistical analy­sis. The sums do not, however, include anonymous attacks.

­ able 2 depicts summary statistics for the number of incidents of transT national terrorism. The summary statistics do not include anonymous attacks where ITERATE reports the nationalities of perpetrators as uncertain or unknown. It also omits incidents whose victims are anonymous.10 The global incident counts range from 0 to 257 (1990–2009) and from 0 to 92 global incidents (2000–2009). Global fatal incidents vary between 0 and 69 (1990– 2009) and 0 and 38 (2000–2009). ­These patterns are congruent with overall patterns of transnational terrorism over time. Since the end of the Cold War, while the volume of transnational terrorism has declined, fatal incidents, both as raw counts and as a proportion of the aggregate count, have risen (Enders and Sandler 2005). In terms of targeted counts, ­t hese have maxima of



Visa Restrictions53

78 (1990–2009) and 44 (2000–2009). Let’s take a look at the countries that generate the greatest number of incidents globally in terms of the nationality of perpetrators.11 The top 95th percentile of the global post-1999 count consists of Af­ghan­i­stan, Colombia, Iraq, Jordan, Nigeria, Pakistan, and Palestine (the West Bank and Gaza).12 In terms of the most terrorism-­ridden dyads, by the nationality-­based coding, Israel-­Palestine, U.S.-­Palestine, U.S.-­Pakistan, ­ ere also U.S.-­Saudi Arabia, and U.S.-­Iraq are at the top. The first three dyads h experience the most fatal incidents. In terms of targeted terrorism post-2000, the United States occupies the top position, followed by the United Kingdom and Israel.

Economic Interdependence Hypotheses 3 and 4 focus on the impact of economic interdependence. To tap into trade interdependence, I create mea­sures that reflect directed dependence. That is, rather than averaging bilateral trade values by dyad, I compute state A’s (the destination state’s) trade dependence on state B (the origin state). This coding aligns with the theoretical intuition that the destination state’s dependence on economic exchange with the origin state ­w ill affect its policies. Thus, directed trade is distinct from both general trade exposure and average bilateral trade. Two operationalization decisions permit me to account for potential reverse causality—­the possibility that visa restrictions impact levels of dyadic trade: first, I take the one period lag of the trade indicators; second, in other models, I utilize the average of values over the 2005–2009 time period.13 Following previous scholarship, I formulate two specifications of trade dependence. The first I label trade salience: this indicator represents the sum of imports from and exports to state A to B as a ratio of state A’s total trade value. Trade salience represents the bilateral trade value as a proportion of the destination state’s commercial portfolio (Mansfield and Pollins 2003). The second indicator expresses the total trade value as a share of the destination country’s GDP. This operationalization taps into the possibility that the impact of trade dependence may be contingent upon the size of the state’s economy (Russett and Oneal 2001). I derive the export and import values from the Correlates of War (COW) trade data (Barbieri and Keshk 2012; Barbieri, Keshk, and Pollins 2009). Version 3 has been extended to include 2009. COW trade data have a dyadic and a monadic version; I make use of the former to create the directed-­dyadic trade mea­sures

54 Chapter 2

and the latter for the country’s total trade. GDP data originate from the World Bank’s World Development Indicators (WDI) (WDI 2015). To capture capital flows, I create a directed capital interdependence indicator that mea­sures FDI divided by the recipient state’s total GDP. Data on dyadic FDI are taken from the OECD’s statistical database (OECD 2009). I complemented missing data from Neumayer’s (2011) replication data. (Neumayer purchased bilateral FDI data from the United Nations’ UNCTAD database.) FDI data are stock foreign investment data. I then perform two transformations. To guard against sensitivity to yearly fluctuations, I average FDI over 2005–2009 and then take the natu­ral log of the values. The latter ensures the results are not influenced by outlier values. Given that the DV captures visa policies for 2010, using the average values also allows me to model the lagged effect of capital dependence. As sensitivity checks, I also deploy overall FDI, incorporated from the World Bank’s WDI.

Control Variables

Next, ­there are two subcategories of control variables included in the models. ­ ese variThe first set pertains to the push-­and-­pull f­actors of migration. Th ables assess host and sending state macroeconomic and institutional traits that drive ­human mobility (Cornelius et al. 2004). The second set gauges dyadic attributes, pertaining to the relationship between sending and receiv­ able 3 enumerates the list of variables and ing states. For the sake of brevity, T respective data sources. A few require elaboration ­here. In terms of the first set, I expect mi­grants to be drawn to demo­cratic states with extensive liberal liberties. Accordingly, I include the revised Polity score for regime type from the Polity IV Proj­ect (Marshall and Jaggers 2013). I use Norris’s (2009) comprehensive democracy data set version 3 to bring in the Freedom House (FH) civil liberties score. I transform the FH score into a dummy equal to 1 for liberal states and 0 other­w ise. I expect negative institutional conditions in ­ oward that end, I also control for origin-­ origin states to affect emigration. T country democracy score. Conflict in the origin state may be positively related to mi­grant outflows; accordingly, I utilize the armed conflict intensity mea­sure from the UCDP/PRIO database, coded 2 if the origin state has seen ­ attle deaths, 1 if it has seen lower-­scale conconflict producing at least 1,000 b flict with at least 25 ­battle deaths, and 0 in the absence of conflict (Gleditsch et al. 2002; Strand, Wilhelmsen, and Gleditsch 2004).



Visa Restrictions55

Distance affects migration patterns much as it affects trade and financial flows (Fitzgerald, Leblang, and Teets 2014). Hence two distance mea­sures, distance between capitals and contiguity, are included. The first originates from Simmons’s (2005) data; contiguity is drawn from the COW Direct Contiguity version 3.1 (Stinnett et al. 2002). Consistent with the work of ­Neumayer (2006) and Avdan (2014a), the models include a battery of shared dyadic traits that tap into links between recipient and sending states: colonial and linguistic ties. ­These are time-­invariant dummies, originating from a variety of source databases that I take from Avdan’s replication data (Avdan 2014a). The inclusion of ­these dummies also controls for pull dynamics whereby mi­grants gravitate ­toward destination countries with a common language or colonial history. The ratio of destination to origin state GDP per capita controls for wealth differentials; we expect mi­g rants to flow from poorer to richer states. Also included is (logged) mi­grant stock in destination states. Mi­g rant networks function as another pull ­factor (Fitzgerald, ­Leblang, and Teets 2014); the same time, traditional emigration states may be more liberal in their visa policies.14 To account for conflict and cooperation dynamics within the dyad, I also inject dummy variables, coded 1 for the existence of an alliance tie and a militarized dispute (MID) within the past de­cade. ­These are coded from the COW Alliance data version 4.1 (COW 2015 Douglas 2009) and COW MID data version 4.0 (Palmer et al. 2015). Th ­ ese indicators are lagged, representing values for 2009 in source data. All other time-­variant control variables are the mean of the 2005–2009 period; exceptions to this rule are listed in ­Table 3. This is consonant with the operationalization of other temporally variant regressors and permits for a check against yearly perturbations in data; it also allows for the lagged effect of in­de­pen­dent variables. Fi­nally, I also control for the EU membership of the destination state. EU harmonization means that we should account for interdependence among member states’ policies; for example, member states have a common list of third-­party countries whose citizens are required to have a visa (Huysmans 2006).15 Visa-­ free mobility within the Eu­ro­pean Union has also been pivotal to deeper integration, suggesting that EU membership may si­mul­ta­neously influence visa freedom and economic ties. To account for this, I include dummies for ­whether the destination and sending states are members of the EU-27, which reflects the Eu­ro­pean Union’s enlargement as of 2007.16 The dependent variable, visa restrictions, is binary, which necessitates the use of a nonlinear estimator such as probit or logit (Greene 1997). The analy­sis

­Table 3. Variable Descriptions and Sources Control Variables Migration Push-­and-­Pull ­Factors

Description of Indicator

Data Source

Origin’s regime type Origin liberal state Origin’s civil liberties

Polity IV score, −10 to 10, ordinal scale 1 if home liberal state, 0 other­w ise, 2006–2009 median 2 to 14 ordinal scale, denotes increasing restriction; dummy below 6 for liberal state. Gross national income per capita, logged, in PPP 0 to 3 scale for increasing levels of civil unrest; 2009 values

Polity IV (Marshall and Jaggers 2013) Norris 2009 Norris 2009

Logged number of mi­grants from origin country Gross domestic product, logged, in PPP 2 to 14 ordinal scale, denotes increasing freedom; 2006–2009 median Polity IV score, −10 to 10, ordinal scale, 2006–2009 median International tourism receipts as % of home GDP

WDI 2015 WDI 2015 Norris 2009

Origin’s GNI per capita Armed po­liti­cal conflict in origin country Recipient’s mi­grant stock Recipient’s GDP per capita Recipient’s civil liberties Recipient’s regime type Recipient’s tourism revenue

WDI 2015 UCDP/PRIO Armed Conflict Database (Gleditsch et al. 2002)

Polity IV (Marshall and Jaggers 2013) WDI 2015

Dyadic Attributes Differences in wealth (GDP per capita differences) Logged distance

Differences in net GDP in PPP

World Bank (WDI), accessed in 2016

Logged distance between capitals in dyad

Simmons 2005 (continued)

­Table 3 (continued) Control Variables Migration Push-­and-­Pull ­Factors

Description of Indicator

Data Source

Dyadic conflict (MID)

1 if militarized interstate dispute existed in 2009

Alliance tie

1 if origin and home joined any type of alliance pact within past 5 years, 0 other­w ise

Common language Colonial link

1 if home and origin share the same language 1 if home and origin belong to same civilization

Correlates of War Proj­ect MID version 4.0 (Palmer et al. 2015) Correlates of War Proj­ect; Alliance Data version 4.1 (Palmer et al. 2015) Avdan 2014a Avdan 2014a

Dyadic Attributes

Supplementary Checks Shared civilization

IGO sum Contiguity (modified direct contiguity score) Bilateral log mi­grant stock

0 other­w ise Modified mea­sure of Huntington’s civilization tie 1 if home and origin belong to same civilization 0 other­w ise Sum of joint intergovernmental organ­ization, coded for past de­cade Modified direct contiguity score Values for 2010

Avdan 2014a Avdan 2014a

COW Direct Contiguity Data 2007 WDI 2015

58 Chapter 2

uses logistic regression, but probit generates equal inferences. Logistic ­regression employs maximum likelihood estimation (MLE) a­ fter linear transformation of the dependent variable into the natu­ral log odds of visa restrictions. In this way, the logistic regression estimates the probability that the recipient state imposes visa restrictions on the origin country. I employ Huber-­W hite robust standard errors clustered by recipient state. This modeling choice accounts for the possibility that destination countries may be uniquely liberal or stringent in visa policies.

Analy­sis ­Table 4 pres­ents the statistical tests for hypotheses 1–4. The first model demonstrates the impact of global terrorism while the second and third models focus on targeted terrorism, using the aforementioned mea­sures based on nationality of victims and venue of attack. Models 4 and 5 focus only on fatal global and targeted attacks, defined as incidents with at least one death. In terms of overall model fit, all models produce significant global Chi statistics (Wald test significant at the 0.000 level). The mean variance inflation ­factor (VIF) is 1.36 with all values well below 5.0, which demonstrates that multicollinearity is not a cause for concern. Across the board, transnational terrorism exerts a positive impact on the probability of visa restrictions, lending support to hypotheses 1 and 2. The coefficient on targeted terrorism in model 3, however, falls short of conventional levels of statistical significance. This indicator counts attacks within destination state by origin nationals. Sensitivity checks show that when incidents are aggregated from 1990 on, the effect is marginally significant (p